JP7605878B2 - Product recommendation method and system - Google Patents

Description

The present invention relates to a product recommendation method and system for recommending appropriate products that reflect the needs of a user.

The dictionary definition of artificial intelligence is a technology that realizes human abilities such as learning, reasoning, perception, and natural language understanding through computer programs. This type of artificial intelligence has made great strides thanks to deep learning.

In particular, with the advancement of artificial intelligence, various language models have been developed, and these language models are now capable of not only recognizing text and understanding its meaning, but also extracting and classifying information from data containing huge amounts of text, such as documents, and even generating text directly.

Such language models are actively used in various fields, such as search engines, document creation (e.g., resume creation, report creation, poster creation, etc.), free dialogue on various topics, data parsing in a given text (e.g., data summarization, classification, etc.), providing expert knowledge, programming, converting a given text into a text of an appropriate style, and many other fields where text-based work can be done.

Based on such language models, it is possible to automate various tasks that were previously performed directly by humans. In particular, it is possible to automate text-based tasks that were previously generated or edited directly by humans in various service fields (e.g., shopping services, news services, etc.). However, in doing so, research is needed into methods of utilizing language models that take into account the characteristics of the service.

The present invention provides a product recommendation method and system that can be used for shopping services.

In particular, the present invention provides a product recommendation method and system that can recommend appropriate products according to the recipient and intent of the gift when a user purchases a product to give as a gift to another person.

Furthermore, the present invention provides a product recommendation method and system that can use a language model to recommend appropriate products depending on the user's situation or the gift-giving situation.

To solve the above problem, the product recommendation method according to the present invention may include a step of identifying target story information including at least one of target subject information and target intention information, a step of acquiring a recommended product category corresponding to the target story information based on sample story information including at least one of sample subject information and sample intention information and sample product categories corresponding to the sample story information, and a step of extracting at least one recommended product that matches the target story information from among products belonging to the recommended product category.

The product recommendation method according to the present invention may include the steps of receiving target story information including information regarding at least one of a gift target and a gift intention from a user terminal, constructing a first prompt to be input to a pre-specified large language model using at least one data set including sample story information and theme information of a sample product category, and the target story information, obtaining a plurality of theme keywords corresponding to the target story information as an output of the large language model based on the first prompt, providing the plurality of theme keywords to the user terminal, and selecting at least one of the plurality of theme keywords in the user terminal, constructing a second prompt including a specific theme keyword selected in the user terminal, and obtaining a recommended gift category corresponding to the target story information from the large language model by processing the second prompt including the specific theme keyword as an input to the large language model.

The product recommendation system according to the present invention may include a storage unit that stores sample story information including at least one of sample subject information and sample intention information, and sample product categories corresponding to the sample story information, and a control unit that performs a process of acquiring a recommended product category corresponding to target story information based on the sample story information and the sample product categories corresponding to the sample story information, using a pre-specified large language model, and a process of extracting at least one recommended product that matches the target story information from among products belonging to the recommended product category.

A program executed by one or more processes in an electronic device according to the present invention and stored on a computer-readable recording medium may include commands to perform the steps of identifying target story information including at least one of target subject information and target intention information, acquiring a recommended product category corresponding to the target story information based on sample story information including at least one of sample subject information and sample intention information and sample product categories corresponding to the sample story information, and extracting at least one recommended product that matches the target story information from among products belonging to the recommended product category.

As described above, the product recommendation method and system according to the present invention can extract recommended products using a large language model that uses prompts containing story information as input.

Here, the story information is configured to include information about the recipient of the recommended product and the intention of the purchaser to purchase the recommended product, so that the present invention can recommend a product that reflects the intention of the user to purchase the recommended product.

In this way, the product recommendation method and system according to the present invention can recommend products based on the story of the user who will give the recommended product, from a large language model, by constructing a prompt using story information that reflects the recipient of the recommended product and the intention of giving the recommended product. Therefore, according to the present invention, it is possible to reduce the time it takes to select a product to give as a gift, and to purchase a product that is optimized for the recipient and the intention of the gift.

1 is a conceptual diagram for explaining an example of application of a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a product recommendation system according to the present invention. 2 is a flowchart illustrating a product recommendation method according to the present invention. 1 is a conceptual diagram for explaining a prompt utilized in a product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a method of inputting story information in a product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining an example of a prompt utilized in the product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining an example of a prompt utilized in the product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining an example of a prompt utilized in the product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining an example of a prompt utilized in the product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining an example of a prompt utilized in the product recommendation method and system according to the present invention; FIG. 2 is a conceptual diagram for explaining product categories and a product recommendation method. FIG. 2 is a conceptual diagram for explaining product categories and a product recommendation method. FIG. 2 is a conceptual diagram for explaining product categories and a product recommendation method. FIG. 2 is a conceptual diagram for explaining a method of extracting recommended products in the product recommendation method and system according to the present invention. FIG. 2 is a conceptual diagram for explaining a method of extracting recommended products in the product recommendation method and system according to the present invention. 1 is a conceptual diagram for explaining a message generating method in a product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining a message generating method in a product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining a message generating method in a product recommendation method and system according to the present invention; 1 is a conceptual diagram for explaining a message generating method in a product recommendation method and system according to the present invention;

Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the attached drawings. However, the same or similar components are denoted by the same or similar reference numerals regardless of the drawing number, and the description thereof will be omitted. The suffixes "module" and "part" of components used in the following description are given or mixed to facilitate the preparation of the specification, and do not have any significance or usefulness in themselves. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of related publicly known technology may obscure the gist of the embodiments disclosed in this specification, the detailed description will be omitted. Note that the attached drawings are merely for the purpose of making the embodiments disclosed in this specification easily understandable, and the technical ideas disclosed in this specification are not limited by the attached drawings, and it should be understood that all modifications, equivalents, or alternatives within the ideas and technical scope of the present invention are included in the present invention.

Terms including ordinal numbers such as first, second, etc. are used to describe various components, but the components are not limited to the terms. The terms are used only to distinguish one component from another.

When a component is said to be "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to the other component, and that there may be additional components in between. In contrast, when a component is said to be "directly connected" or "directly connected" to another component, it should be understood that there are no additional components in between.

Singular expressions include plural expressions unless otherwise specified.

In this specification, the terms "comprise" and "have" are intended to specify the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but should be understood not to preclude the presence or additional possibility of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

As mentioned above, various language models have been developed with the advancement of artificial intelligence, and based on these, it has become possible to automate various tasks that were previously performed directly by humans. In particular, it has become possible to automate various tasks that were previously selected or generated directly by humans in various service fields (e.g., shopping services, news services, etc.). In particular, the present invention provides a product recommendation method and system that can recommend products suitable for a user by taking into account the various circumstances of the user who purchases the product.

Users purchase products over the Internet for themselves or others, and especially when purchasing products for others, there may be a recipient and intention to give the purchased product.

For example, purchasing a product for someone else may result in giving it as a gift to that person. In this case, the user can search for and purchase the product on the Internet, keeping in mind the intention and recipient of the gift.

The product recommendation method and system according to the present invention are intended to facilitate the convenience of users who purchase products for others. The present invention can provide a product recommendation method and system that can reduce the time a user spends selecting products to purchase for others and can recommend products according to the user's product purchase intention and the recipient of the product.

In the present invention, for the sake of convenience, the terms "purchase of a product" and "purchase of a gift" are used interchangeably, and the present invention is not necessarily limited to purchasing a gift. In other words, it goes without saying that the present invention can be used when a user purchases a product for himself or herself, or when purchasing a product for someone else who is not intended as a gift. Therefore, in the present invention, "purchase of a gift" may be replaced with "purchase of a product" or may be interpreted as "purchase of a product". However, in the case of purchasing a gift, there is a difference in that the purchaser of the product and the user of the product are different, and therefore the product cannot be recommended in a way that directly reflects the purchaser's behavioral characteristics and preferences. The present invention provides a method for solving such a problem. In this case, the term "product" may be used not only for tangible goods but also for intangible services.

More specifically, as shown in FIG. 1, the product recommendation method and system according to the present invention may use story information to recommend products to a user. Here, the products recommended to the user may be products that the user can purchase on the Internet.

Story information in the present invention may include information on at least one of the target and the intention. Here, the "target" may be understood as the party who receives the product recommended in the product recommendation system of the present invention. For example, when a user uses the product recommendation system of the present invention for a gift, the "target" may be the person who receives the gift (the party or the other person). Here, the information on the target (or target information) included in the story information may be composed of at least one of various pieces of information that can describe a specific person corresponding to the target, such as gender, age, occupation, relationship with the user, hobbies, preferences, and place of residence.

Furthermore, the "intention" may be interpreted as the reason or purpose of a user who purchases a product recommended in the product recommendation system according to the present invention. For example, if a user purchases a product as a gift, the "intention" may be interpreted as at least one of various pieces of information that may explain the intention of the gift (e.g., a housewarming gift, a moving gift, a wedding anniversary gift, etc.).

On the other hand, in the product recommendation method and system according to the present invention, the user can input story information in various ways. For example, the user can input story information on a service page 110 provided by the product recommendation system according to the present invention, or the story information can be selected from a list of story information provided 112, as shown in FIG. 1(a).

On the other hand, in the product recommendation system according to the present invention, a language model (LM) or a large language model 10 may be used to obtain a recommended product category corresponding to the story information. Furthermore, in the product recommendation system according to the present invention, based on this, as shown in FIG. 1(b), recommended products corresponding to the story information may be extracted and provided to the user. Also, at least one keyword related to the recommended products or keywords that can classify them may be extracted and provided to the user.

On the other hand, the language model utilized in the present invention is not limited in type, and may be any language model that has been trained on various information. On the other hand, a large-scale language model or a super-large language model, which has been actively researched in recent years, is a model trained with a large amount of data using a model having more than 10 times as many parameters (or parameters) (e.g., more than 100 billion parameters) as compared to a conventional general language model, and can be inferred without fine-tuning the model using a method such as few-shot learning. In such a large-scale language model, a pre-trained model can understand patterns in the prompt by inputting a prompt including a small amount of example data format, and can generate and output a probability distribution of a word sequence or a word (or sentence) based on the same.

Here, the recommended product category may be interpreted as the category or type to which the recommended product belongs. Product categories may be classified based on various criteria such as product use, product color, and product price range. For example, product categories may include categories such as accessories, stationery, and automotive supplies depending on the product use. Here, the recommended product in the present invention may mean a product that belongs to a specific category and is actually available for purchase on the Internet. In other words, the product recommendation system according to the present invention may recommend products to the user and even support the user in purchasing the recommended product.

Here, the recommended product may be made available for purchase on a shopping service page 120 (e.g., a web page or an application page) provided by a shopping service linked to the product recommendation system according to the present invention, as shown in FIG. 1(b).

Hereinafter, the product recommendation method and system according to the present invention will be described in more detail with reference to the accompanying drawings. FIG. 2 is a conceptual diagram for explaining the product recommendation system according to the present invention, and FIG. 3 is a flow chart for explaining the product recommendation method according to the present invention. FIG. 4 is a conceptual diagram for explaining a prompt used in the product recommendation method and system according to the present invention, and FIGS. 5, 6, 7A, 7B, 8, 9A, and 9B are conceptual diagrams for explaining a method of inputting story information in the product recommendation method and system according to the present invention. Also, FIGS. 10A, 10B, 11, 12A, and 12B are conceptual diagrams for explaining examples of prompts used in the product recommendation method and system according to the present invention, and FIGS. 13A, 13B, and 13C are conceptual diagrams for explaining product categories and a product recommendation method. FIGS. 14 and 15 are conceptual diagrams for explaining a method of extracting recommended products in the product recommendation method and system according to the present invention, and FIGS. 16A, 16B, 17A, and 17B are conceptual diagrams for explaining a method of generating a message in the product recommendation method and system according to the present invention.

First, as shown in FIG. 2, the product recommendation system 200 according to the present invention may include at least one of an input unit 210, a storage unit 220, and a control unit 230. Although not shown, the product recommendation system 200 may further include a communication unit, which may be configured to communicate with at least one server, electronic device (or user terminal).

Furthermore, the product recommendation system 200 may be configured to include at least one of the large language model 10 and the product recommendation engine 20, and may be configured to operate in conjunction with at least one of them.

The product recommendation system 200 according to the present invention may acquire a recommended product category from the large language model 10 based on story information input to the large language model 10, and extract recommended products corresponding to the recommended product category using the product recommendation engine 20. At least a part of the configuration of the product recommendation system 200 may be configured to process information input to at least one of the large language model 10 and the product recommendation engine 20, or to process information output from at least one of the large language model 10 and the product recommendation engine 20.

Regarding the input unit 210, the input unit 210 may serve to input information constituting a prompt to be input to the large language model 10.

The input unit 210 may include i) an interface that connects a user (or a user terminal) to the product recommendation system 200 according to the present invention, or ii) an interface that connects the product recommendation system 200 according to the present invention to at least one other system. Furthermore, the input unit 210 may include an interface that connects between components included in the product recommendation system 200 according to the present invention.

In the present invention, by inputting a prompt into the large language model 10, the result may be a recommended product category and the corresponding recommended products.

Here, the recommended product category may be obtained from the large language model 10, and the recommended products belonging to the recommended product category may be obtained from the product recommendation engine 20. The control unit 230 may control a series of processes for obtaining recommended products from the product recommendation engine 20 by inputting information related to the recommended product category obtained from the large language model 10 to the product recommendation engine 20.

The prompt input to the large language model 10 is data input to the large language model 10, as shown in (a) of FIG. 4, and may include datasets 411, 412, 413 (hereinafter referred to as "sample datasets") that are referenced in the large language model 10, and a dataset 414 (hereinafter referred to as "target dataset") that corresponds to the information to be obtained from the large language model 10.

Such data sets 411, 412, 413, 414 may include the story information described above.

The sample dataset may be configured with a pair of sample story information and information on the corresponding sample product category. Here, the sample story information may be generated using information created or produced by the user, or may be generated by extracting the target and intent from the user's product purchase details. In addition, the sample story information may be generated by inputting a prompt including the above-mentioned information (e.g., information created or produced by the user, the user's product purchase details, etc.) into a trained language model or a large language model.

The target story information included in the target dataset may be data input in text format by the user, as described with reference to FIG. 1(a), or may be data selected from a story information list. In addition, the target story information may be generated by combining at least one of various pieces of information such as basic information about the user (e.g., age, gender, etc.) and information input by the user (e.g., target, intent, etc.).

The target dataset does not include a target product category, i.e., a recommended product category, that corresponds to the target story information, which may be configured to be obtained from the large language model 10.

The prompt 410 may be configured to include at least one of the sample data sets 411, 412, 413 and at least one of the target data set 414. The configuration of the prompt is described in more detail below. The configuration of the information and the method of obtaining the corresponding product category are also described in more detail below.

Meanwhile, according to another embodiment of the present invention, the prompt input to the large language model 10 may be an embedding vector generated using a trained embedding model instead of a text format. Here, the prompt embedding model may be trained using sample story information created or generated by the user, or sample story information generated by extracting the target and intent from the user's product purchase details.

Next, various information may be stored in the storage unit 220. The storage unit 220 may exist at least one inside or outside the product recommendation system according to the present invention. It goes without saying that the storage unit 220 may be configured as a server or as a cloud-based storage.

The storage unit 220 may store data relating to at least one of the multiple sample datasets and the target dataset.

The multiple sample data sets may be configured to include at least one of story information, product category information, theme information, keyword information, and tag information.

At least a portion of the various information contained in the sample data sets listed above may be comprised of information collected at various pages accessible via the Internet.

The collected information may consist of information created or produced by any user on the Internet. Such data may be called UGC (User Generated Content). Furthermore, at least a portion of the various information contained in the sample data set may consist of information generated by the large language model 10.

In this manner, the storage unit 220 may store at least one of story information, product category information, theme information, keyword information, and tag information collected through various routes, and these pieces of information may be paired to form a data set that is input to the prompt.

Next, the control unit 230 may perform the overall control required to obtain the recommended product category and the recommended product. As an example, the control unit 230 may be configured to work in conjunction with the large language model 10 so that the prompt 410 is processed as an input to the large language model 10. Furthermore, the product recommendation system 200 according to the present invention may include the large language model 10 as one component, and in this case, the control unit 230 may perform appropriate control over the large language model 10.

As another example, the control unit 230 may be configured to work in conjunction with the product recommendation engine 20 so that information regarding recommended product categories obtained from the large language model 10 is processed as input to the product recommendation engine 20.

Meanwhile, the information generated by the large language model 10 may be stored in the storage unit 220 or may be configured to be output from an output unit (not shown) so that the user can check it, and is processed appropriately according to the situation. The control unit 230 may process the information generated by the large language model 10, and may provide the information to the user terminal, for example, when information regarding recommended categories needs to be visually output to the user.

In the product recommendation method and system according to the present invention, as shown in Figures 1 and 2, a data set including at least one of story information and information related to the corresponding product category may be input to a large language model 10 to obtain a recommended product category.

As described with reference to FIG. 4, the large language model 10 is input with a prompt 410 consisting of sample data sets 411, 412, 413 and a target data set 414, and each data set may consist of data fields corresponding to a type of data (category, data type, data attribute) and corresponding data values.

As shown in FIG. 4(a), a data field may be interpreted as defining the meaning of a data value or the type of the data value, such as "story information" or "product category." Such a data field may also be interpreted as "category name," "data type," "data attribute," etc.

On the other hand, data fields may follow predefined rules so that they can be recognized by the large language model 10. For example, the large language model 10 may recognize that data entered after "Story information:" is data that corresponds to the data field "Story information".

The large language model 10 may refer to data pairs contained in a dataset to generate data corresponding to a specific data field (e.g., "Product Category:") contained in the target dataset.

In the present invention, as shown by reference numeral 414b in FIG. 4A, the prompt may be configured so that the end of 410 is composed of only a data field, and the corresponding data value is obtained as the output of the large language model 10. As a result, as shown by reference numeral 424b in FIG. 4B, a data value (picture frame, potted plant, vase, diffuser, candle warmer, lighting, etc.) for the "product category" data field can be obtained.

It goes without saying that you may change the names of the data fields to more appropriate terms depending on the situation.

Next, the product recommendation engine 20 is an engine that recommends products that belong to a recommended product category obtained from the large language model 10, and may be configured to recommend at least one appropriate product based on the similarity between at least two of the story information, the product category, and the product name. In the present invention, the product recommendation engine 20 may be of any type.

Below, we will explain in more detail how to generate product recommendation information with reference to the product recommendation system 200 according to the present invention described above.

As shown in FIG. 3, in the present invention, a process of identifying target story information may be performed (S310).

As mentioned above, the target story information may refer to the product category and story information that is the subject of the product recommendation.

The target story information may include information on at least one of the target subject and the target intent. Here, the target subject may be understood as the party who will receive the product recommended in the product recommendation system according to the present invention. For example, when a user uses the product recommendation system according to the present invention for a gift, the "target subject" is the person (party or other person) who will receive the gift. Here, the information on the target subject (or subject information) included in the target story information may be composed of at least one of various pieces of information that can describe a specific person who corresponds to the target subject, such as gender, age, occupation, relationship with the user, hobbies, preferences, and place of residence.

Furthermore, the "target intent" may be interpreted as the reason or purpose of a user who purchases a product recommended in the product recommendation system according to the present invention. For example, if a user purchases a product as a gift, the "target intent" may be interpreted as at least one of various pieces of information that may explain the intent of the gift (e.g., a housewarming gift, a moving gift, a wedding anniversary gift, etc.).

Meanwhile, the aforementioned "target story information," "target object," and "target intent" have the same or similar meanings as the "story information," "object," and "intent" defined above, but for ease of explanation, the term "target" has been added to distinguish them from "sample story information," "sample object information," and "sample intent information."

The control unit 230 may identify (or generate) target story information based on at least one of information input from a user terminal (e.g., user input information) and user information linked to a user account.

The control unit 230 may further include user information about the user in the target story information to identify (or generate) the target story information.

The control unit 230 may extract specific user information to be included in the target story information from user information linked to the user account related to the user, or may request the user to input the specific user information. Here, the specific user information may include information regarding at least one of the user's age and gender.

Here, the user's age may be used to infer the generation of the person who will receive the product (e.g., the person who will receive a gift), and the user's gender may be used to infer the purchasing popularity of the product (e.g., a gift) among users who purchase the product.

For example, if the target story information includes information about age (e.g., a teenage user) as user information, the teenage user's friends may be inferred to be teenagers of the same age as the user, and the teenage user's parents may be inferred to be in their 40s to 50s. In this way, if the target story information includes age as user information, the age of the person who will receive the gift may be estimated.

Furthermore, the gender of the user may be used to consider the popularity of gifts for users of that gender. For example, women in their twenties tend to give accessories as birthday presents to their younger sisters, while men in their twenties tend to give digital devices as birthday presents to their younger brothers.

In the present invention, by including user information in the target story information, the age of the person receiving the gift may be inferred, and product categories and products may be recommended that take into account the popularity of gifts for the user giving the gift.

On the other hand, the target story information may be included in the prompt 410 so as to be processed as an input to the large language model 10 described above. The control unit 230 may generate the target story information using at least one of the user input information and the user information based on natural language processing (NLP). The control unit 230 may generate the target story information in a sentence format so as to include at least one of the user input information and the user information. Therefore, the user input information and the user information are not necessarily limited to a sentence format structure, and may be processed by the control unit 230 to have a data structure in a form that can be processed in the large language model 10.

However, in the present invention, information may be input by the user in a variety of ways and used to identify target story information.

As an example, as shown in FIG. 5, in the present invention, target story information may be identified based on information entered into an information input area 510 included in a specific page 500 provided by a product recommendation service or shopping service.

The information input area 510 may be composed of various information input areas, such as a product search area or a story information input area.

The control unit 230 may identify target story information using user input information (e.g., "My best friend moved last week. What should I buy for her? Something expensive is fine.") input from the user terminal (or user account) in the information input area 510.

In this manner, the control unit 230 may identify target story information based on a product recommendation request received in the information input area 510 from a user account.

The control unit 230 may extract information on at least one of the target object and the target intent from the user input information. The control unit 230 may analyze the user input information based on natural language processing, and extract information on at least one of the target object (e.g., a best friend) and the target intent (e.g., a housewarming gift) necessary to identify the target story information. The extracted information may then be used to generate the target story information.

As another example, the control unit 230 may identify the target story information using a chatbot, as shown in (a) and (b) of FIG. 6. The control unit 230 may provide messages 611, 612, and 616 for identifying the target story information to the user terminal. The control unit 230 may also identify the target story information based on the message 616 received from the user terminal.

The dialogue messages 611, 612, 616 for identifying the target story information may be provided to the user terminal multiple times, and the content of the next message may be appropriately selected and provided to the user terminal depending on the content of the message 616 received from the user terminal. Here, the control unit 230 may identify information that is missing from the configuration of the target story information, and may configure a message to be provided to the user in a form that further guides that information.

On the other hand, in the product recommendation system 200 according to the present invention, a chatbot may be used to provide information (e.g., a question message) for identifying target story information to a user terminal, and at least one of the extracted information on the recommended product category and information on the recommended product 614, 616, 623 may be provided to the user terminal based on the target story information by the chatbot. This will be described in detail later.

As yet another example, as shown in Figures 7A and 7B, in the product recommendation system 200 according to the present invention, the control unit 230 may receive user input information for identifying target story information on a specific page provided to the user terminal.

As shown in FIG. 7A, the specific page 710 may be configured to include at least one of a first input area 711 for receiving target information corresponding to a recipient (or other person, e.g., a person receiving a gift) who will receive the product recommended in the product recommendation system 200 according to the present invention, and a second input area 712 for receiving target intention information corresponding to an intention to give the recommended product to the recipient.

That is, the specific page 710 may include a plurality of input areas 711, 712 for receiving from the user the recipient (target) who will receive the recommended product and the intention to give (present) the product. The control unit 230 may identify the story information based on the user input information received in the plurality of input areas 711, 712.

On the other hand, as shown in Figures 7A and 7B, the information 713, 715 constituting the specific page 710 may be changed according to user input information entered on the specific page 710.

For example, as shown in FIG. 7A, in the specific page 710, the target and the intention may be input in order by the user. Thus, information on a plurality of different targets (friends, parents, baby, mother, sister, etc.) may be provided in one area 713 of the specific page 710. Here, the information on the target provided to the user may be appropriately determined taking into consideration the user information (e.g., gender, age, other family relationship information, etc.). For example, if the user is a woman in her 40s, the targets may include her husband, parents-in-law, children, etc., and if the user is a teenage male, the targets may include friends, teachers, parents, etc.

Such information about a target may be represented by a graphic object (or icon) as shown in FIG. 7A, and when one of the graphic objects (e.g., graphic object 713a corresponding to "friends") is selected, the control unit 230 may identify information matching the selected graphic object (e.g., "friends") as the target.

Information about a particular target may be input into a first input area 711, as shown in FIG. 7B. Furthermore, it goes without saying that a user may directly input information about the target into the first input area 711. That is, the first input area 711 may be configured to directly input information received from a user terminal.

Next, as shown in FIG. 7B, an area 715 of a particular page 710 may be provided with information about a number of different intentions (satisfaction, gratitude, birthday, friendship, housewarming, etc.). Here, the intentions provided to the user may be determined taking into consideration information about a preselected target and user information. For example, when the target is different, such as when the target is a "friend" and when the target is a "teacher," the information about the intentions provided may also be different.

Such information regarding the intention may be represented by a graphic object (or icon) as shown in FIG. 7A, and when one of the graphic objects (e.g., graphic object 715a corresponding to "housewarming gift") is selected, the control unit 230 may identify the information that matches the selected graphic object (e.g., "housewarming gift") as the intention.

In addition, information regarding a specific intention may be input into the second input area 712. The second input area may be configured to display intentions having the same characteristics among a plurality of intentions together, and one of the intentions provided together may be selected by the user. Furthermore, it goes without saying that the user may directly input information regarding the intention into the second input area 712.

In this way, in the present invention, information regarding the target and intent may be received based on user input information. When the control unit 230 receives the information regarding the target and intent, the control unit 230 may use the received information to identify story information. In other words, the target target and target intent may be identified by the method described above, and target story information may be identified based on the identified information.

The control unit 230 may use a natural language processing algorithm to generate sentences containing information about the object (or target object) and intent (or target intent) received in the first input area 711 and the second input area 712, as shown in Figures 7A and 7B. Such sentences may be processed into a form that is included in a prompt of the large language model 10, as described above.

As yet another example, the control unit 230 may use a specific page 800 as shown in FIG. 8 to provide a list 810 including various story information to the user terminal, and the user may identify the story information by selecting one of the provided story information.

As shown in the figure, the specific page 800 may include a list 810 consisting of multiple different pieces of story information. When any item 820 in the list 810 is selected on the user terminal, the control unit 230 may identify the story information corresponding to the selected item (e.g., "Please recommend a gift for the meeting with my future mother-in-law.") as the target story information.

As yet another example, the control unit 230 may allow the user to input story information using a specific page 910, as shown in FIG. 9A. As shown in the figure, the specific page 910 may be provided with guide information 911 for guiding the user in inputting the story information.

Furthermore, the specific page 910 may be provided with guide information 912 to guide the input of information that must be included in the story information (e.g., information about the subject and intent). By referring to the information, the user can input the story information including the subject and intent.

Additionally, the particular page 910 may be provided with sample story information 913 that includes information about the subject and intent. The sample story information allows the user to intuitively recognize what story information needs to be entered.

On the other hand, the sample story information may be configured to be selected by a user terminal. By selecting the sample story information, the user can input story information.

Furthermore, as shown in FIG. 9B, the specific page 920 may be provided with an additional information input area 922 for more specifically specifying the story information.

The control unit 230 may provide an information input (or selection) area 922a on the specific page 920 for more specifically identifying at least one of the subject and intent that must be included in the story information.

Furthermore, the specific page 920 may be provided with an area 922b in which various information required for product recommendations, such as the price range of the product, can be entered.

The control unit 230 may identify story information based on user input information received from a user terminal on such a page.

In this way, in the product recommendation system 200 according to the present invention, story information (target story information) may be identified based on user input information received by various methods.

As mentioned above, the user input information does not have to be in the form of a sentence. In that case, the control unit 230 may use the user input information to generate sentence 1014a in a form that can be included in the prompt 1010, as shown in FIG. 10A (a).

For example, if the user input information consists of multiple words, the control unit 230 may combine the multiple words to generate a sentence and include it in a prompt that is processed as input to the large language model 10.

On the other hand, the control unit 230 may specify story information using user information stored in a specific server in association with the user account in order to recommend a product more suitable for the user. In this case, the control unit 230 may specify story information 1045 as shown in FIG. 10B (b) by further utilizing user information 1032 (e.g., male, in his teens) in addition to story information 1031 based on information input by the user (user input information) as shown in FIG. 10B (a). The story information may include first information 1045a based on the user information (e.g., "I am a male student in his teens.") and second information 1045b based on the user input information ("I am a woman in my 30s who likes interior design. Please recommend a gift for a housewarming."). The control unit 230 may generate at least one sentence so that the user input information and the user information are included, and include the sentence in the prompt 1040 that is processed as an input to the large language model 10.

On the other hand, the user information may include at least one of the gender, age, residential area, and occupation of a specific user corresponding to a specific user account, and may also include various other information stored in the storage unit 220 regarding the specific user.

In the present invention, when recommending products, user information may be taken into consideration to recommend product categories and products that suit the individual circumstances of each user. For example, if age is input as user information, the price range of products that can be purchased may differ between a user in his/her teens and a user in his/her thirties.

Furthermore, in the present invention, when recommending a product, by taking user information into consideration, information about the recipient of the product (e.g., the person who will receive a gift corresponding to the product) may be inferred, and at least one of a product category and a product suitable for the recipient may be recommended using the inferred information about the recipient. For example, age (e.g., a teenage user) may be input as user information, and user input information (e.g., "I'm thinking of giving a gift to a friend") may be received. In this case, the teenage user's friends are inferred to be teenagers of the same age as the user, and as a result, the recipient of the gift is inferred (or estimated) to be a teenage user. In this way, in the present invention, a product category or product suitable for the recipient can be recommended by inferring information about the recipient based on user information.

As yet another example, the user's age (e.g., a teenage user) may be input as user information, and user input information (e.g., "I'm thinking of giving a present to my parents") may be received. In this case, the parents of a teenage user are statistically inferred to be in their 40s to 50s, and as a result, the recipient of the present is inferred (or estimated) to be a user in their 40s to 50s. In this way, in the present invention, by inferring information about the recipient based on the user information, it is possible to recommend a product category or product suitable for the recipient (e.g., a user in their 40s to 50s).

In this way, in the present invention, story information can be identified that reflects user information so that product categories and products appropriate to each user's situation are recommended.

As a result, in the present invention, the type of recommended product category obtained as the output of the large language model 10 may vary depending on the user information. Similarly, the type of recommended product obtained as the output of the product recommendation engine 20 may vary depending on the user information.

As described above, in the present invention, information including at least one of a target and an intent is input by a user in various ways, and story information can be identified based on the information.

When story information is identified in this manner, i.e., when target story information is identified, in the present invention, a process of obtaining a recommended product category corresponding to the target story information based on the sample story information and the sample product category may be performed (S320).

The control unit 230 may configure a prompt to include target story information in order to obtain a recommended product category. The control unit 230 may configure a prompt to include sample story information, a sample product category corresponding to the sample story information, and the target story information. Furthermore, the control unit 230 may process the prompt as an input to the large language model 10.

Furthermore, in the large language model 10, based on the input prompt, a product recommendation category may be generated as an output (or response) to the prompt.

As described with reference to FIG. 4, the prompt may include sample data sets 411, 412, 413 that are referenced in the large language model 10, and a target data set 414 that corresponds to information to be obtained from the large language model 10.

The types of data fields included in sample data sets 411, 412, and 413 may be varied.

As an example, sample data sets 411, 412, and 413 may include a data field corresponding to story information and a data field corresponding to product category, as described with reference to FIG. 4.

As another example, the sample data set may include a data field corresponding to story information, a data field corresponding to tags, and a data field corresponding to product categories. Here, tags may be referred to as images, topics, keywords, etc.

In the large language model 10, the recommended product category may be output by referring to the data constituting the sample dataset included in the prompt and the data constituting the target dataset.

On the other hand, the control unit 230 may select information to be input in the prompt, such as data constituting a sample data set, for example, sample story information, sample tag information, sample product category, etc. This information may be stored and exist in the storage unit 220.

These pieces of information may also be based on UGC collected on the Internet. Furthermore, these pieces of information may be generated from a large language model 10.

The control unit 230 may configure a prompt 1110 to obtain sample story information and input it to the large language model, as shown in FIG. 11(a). Here, as shown in FIG. 11(a), a data field corresponding to the story information of the target data set 1111 included in the prompt 1110 may be left empty.

In the large language model 10, at least one piece of story information 1120 may be acquired based on a prompt in which a data set (or a sample data set) consisting of a gift category and a story information data field is input, as shown in FIG. 11(b). Various pieces of story information acquired in this manner may be utilized as information included in the sample data set, as described with reference to FIG. 4. Meanwhile, information in the data fields corresponding to the gift category and story information included in the sample data set shown in FIG. 11(a) may be information collected on the Internet. This information may be UGC generated by a user.

In this way, the control unit 230 may obtain a recommended product category from the large language model 10 as an output (response) to a prompt based on the prompt input to the large language model 10.

On the other hand, when a recommended product category is obtained from the large language model 10, the control unit 230 may provide information related to the recommended product category to the user terminal. For example, as shown in FIG. 13A, information related to the recommended product category obtained from the large language model 10 (e.g., kitchen supplies, cleaning supplies, interior goods, bedding) may be provided in an area 1330 of a specific page 1310 where story information is input. The user can select the product category they wish to purchase by looking at the recommended product categories and selecting a specific recommended product category (e.g., "interior goods").

In this way, the product categories obtained from the large language model 10 may be provided to the user terminal on a page where story information (user input information) is input from the user terminal, as a response to the user input information. Thus, the user can predict for himself what products will be recommended for the input information, and can specify the range of recommended products by selecting a specific product category.

As an example, as shown in FIG. 5, information 530 regarding recommended product categories may be provided in response to information entered in input area 510.

As another example, as shown in FIG. 6(a), the chatbot may provide information 615 about recommended product categories in response to a message 613 entered by the user.

So far, we have explained a method for immediately obtaining information about recommended product categories from the large language model 10 in response to story information based on user input information entered by a user.

Furthermore, the control unit 230 may be configured to primarily acquire tags corresponding to story information from the large language model 10 in order to reflect the user's intention when recommending a product. Then, when at least one of the acquired tags is selected by the user, the control unit 230 may be configured to secondarily acquire a recommended product category that reflects information related to the selected tag from the large language model 10. In this way, the control unit 230 can recommend products that reflect the story information based on the user input information and the user's intention.

More specifically, the control unit 230 may receive target story information including information on at least one of a target object (e.g., a gift object) and a target intent (e.g., a gift intent) from the user terminal. Also, at least one data set (or sample data set) including sample story information and a theme keyword (or tag) of a sample product category, and the target story information may be used to configure a first prompt to be input to a pre-specified large language model. As shown in FIG. 12A, a prompt 1210 consisting of a sample data set including sample story information and sample tags, and a target data set 1220 including target story information and tag data fields may be input to the large language model 10. The control unit 230 may obtain a plurality of tags (or theme keywords) corresponding to the target story information as an output 1210 of the large language model based on the prompt 1210.

As shown in FIG. 6 and FIG. 13B, the acquired tags (e.g., luxury, health essentials, great taste, great atmosphere, Instagram-worthy, home cafe, etc.) may be provided in an area 615, 1360 on a particular page 610, 1350 in response to user input information.

Furthermore, in the present invention, a plurality of tags (or theme keywords) may be provided to the user terminal, and at least one of the plurality of tags may be selected on the user terminal. Furthermore, the control unit 230 may configure the prompt (or the second prompt) to include the selected tag, and may obtain a recommended product category from the second prompt.

The control unit 230 may process the second prompt including the specific tag selected by the user as an input to the large language model 10 to obtain a recommended gift category corresponding to the target story information from the large language model 10.

In this way, the control unit 230 may obtain a recommended product category that reflects the user's intention by sequentially inputting multiple prompts into the large language model 10. On the other hand, at least one data field constituting the first prompt and one data field constituting the second prompt may be different.

For example, the first prompt may include a data field corresponding to story information and a data field corresponding to tags, and the second prompt may include a data field corresponding to story information, a data field corresponding to tags, and a data field corresponding to a product category.

Here, the second prompt may be composed of at least one first data set including sample story information, tag information (e.g., theme information of a sample product category) and a sample product category, and a second data set including target story information and specific tags. Here, the specific tags constituting the second data set may be composed of tags selected on the user terminal.

Meanwhile, once the recommended product category is obtained using the various methods described above, the product recommendation system and method according to the present invention may then perform a process of obtaining recommended products corresponding to the recommended product category (S330) and providing recommended product information (S340).

When a recommended product category is obtained from the large language model 10, the control unit 230 may extract at least one recommended product corresponding to the obtained recommended product category from the product recommendation engine 20.

The extracted recommended products may be comprised of products available for purchase on the Internet, and information 540, 550, 616, 623, 714, and 716 relating to the recommended products may be provided on the page that received the user input information, as shown in Figures 5, 6, 7A, and 7B.

More specifically, the recommended products may be configured to be available for purchase at least one product sales site accessible via the Internet, and when the recommended products are extracted, product sales information linked to the sales page of the recommended products may be provided to the user terminal.

On the other hand, the control unit 230 may provide information about products belonging to a recommended product category, the specific product category of which is obtained from the user based on at least one of the user input information and the user selected tag, in an area 1340 of the specific page 1310, as shown in FIG. 13A.

As shown in FIG. 13A, when a user selects one of a plurality of recommended product categories (e.g., "interior goods"), the control unit 230 may provide the user terminal with product information 1340 belonging to the selected recommended product category. The control unit 230 may control the product recommendation engine 20 so that products corresponding to the recommended product category are extracted from the product recommendation engine 20.

Furthermore, as shown in FIG. 13C, when one of the multiple tags acquired from the user (e.g., "Great atmosphere" or "Instagrammable") is selected, the control unit 230 may provide the product information 1395 belonging to the selected tag to the user terminal. The control unit 230 may control the product recommendation engine 20 so that products corresponding to the selected tag are extracted from the product recommendation engine 20.

On the other hand, the control unit 230 may extract at least one of the multiple products as a recommended product in descending order of similarity based on the similarity between the target story information and multiple products belonging to a recommended product category.

As shown in FIG. 14, the control unit 230 may determine the similarity 1410 between the user input information (or story information 1421) and the product name 1431 extracted from the product recommendation engine based on the embedded model structure, and extract the top n products with the highest similarity as recommended products as shown in (a), (b), and (c) of FIG. 15, and provide the user terminal with product information (e.g., thumbnails) linked to the sales pages of the recommended products.

As shown in FIG. 14, the embedding model may embed user input information and product names using values from hidden layers of large language models (e.g., Hyperclova) 1422, 1432.

Furthermore, the present invention may further provide a message (e.g., gift message) creation service that corresponds to story information based on user input information.

For example, a recommended product may be provided to the user, and the recommended product may be purchased on the user terminal. Thus, the user may purchase the recommended product as a gift corresponding to the story information input as described above. There may also be a need for the user to create a gift message according to the story information and the recommended product that has been purchased.

When a recommended product is purchased, for example, when a product is purchased on a product sales page included in the product information related to the recommended product, the control unit 230 may generate a message related to at least one of the story information and the recommended product using the story information and information related to the recommended product.

Here, the control unit 230 may generate a message such that the target information included in the story information is included as information regarding the recipient to whom the recommended product is to be gifted, and the intention information included in the story information is included as information regarding the intention to gift the recommended product.

Furthermore, the control unit 230 may use the large language model 10 to generate a present message 1600 as shown in Figures 16A and 16B.

As an example, as shown in FIG. 16A, a prompt including a story (or story information) 1610 and an ordered item (or purchased item) 1620 may be input to the large language model 10, and a gift message (e.g., a sample message for a card) 1630 may be obtained in response thereto.

The control unit 230 may identify a sample data set that includes data fields and corresponding data values corresponding to the story information, the ordered items, and the gift message, respectively, to obtain the gift message from the large language model 10. The prompt may also be configured to include the sample data set.

The prompt may further include a target data set including data fields corresponding to the target present message to be generated, and the target data set may include target story information, data fields and corresponding data values corresponding to the target order items, and data fields corresponding to the target present message. Here, the data values of the data fields corresponding to the target present message may exist as empty fields.

In this way, the control unit 230 may configure a prompt to generate a gift message based on user input information input by the user and the user's product purchase history, and input the prompt to the large language model 10. The control unit 230 may also obtain the gift message from the large language model 10.

As another example, as shown in FIG. 16B, the prompt 1640 input to the large language model may include data fields corresponding to the intent (e.g., intent to give a gift or intent to purchase a product), the giver (e.g., the person purchasing the product or the person giving the gift), the recipient (e.g., the person receiving the product or the person receiving the gift), the date (e.g., the date of purchase of the product or the date of giving the gift), the gift (e.g., the type of product purchased, the product name, etc.), and the card (e.g., the gift message).

As shown in the figure, the prompt 1640 may include a sample data set that includes the data fields and their corresponding data values described above. The control unit 230 may use the large language model 10 to obtain a gift message (e.g., corresponding to the "card" data field) included in the target data set 1650.

In this manner, the control unit 230 may identify a sample data set that includes data fields and corresponding data values for the intent, giver, recipient, date, gift, and card. The prompt may also be configured to include the sample data set.

As shown in FIG. 16B, the prompt 1640 may further include a target data set 1650 including data fields corresponding to the target present message (e.g., a card) to be generated, and the target data set may include data fields corresponding to the intent, the giver, the recipient, the date, and the present, and corresponding data values, and a data field corresponding to the target present message (e.g., a card). Here, the data value of the data field corresponding to the target present message may exist as an empty field.

In this manner, the control unit 230 may configure a prompt to generate a present message and input the prompt to the large language model 10. The control unit 230 may also obtain the present message from the large language model 10.

On the other hand, gift messages 1731, 1732 obtained based on a large language model or obtained by other methods may be provided in the message input area 1730 of the gift message creation page 1700, as shown in Figures 17A and 17B.

The control unit 230 may create a gift message based on at least one of the story information, the purchased product, and the user information.

The content and tone of the gift message may be determined based on at least one of the story information, the purchased product, and the user information.

The control unit 230 may generate a gift message taking into consideration the target (the person to whom the gift is to be given), user information, the user's tone of voice, frequency of stamp use, and the like. For example, the control unit 230 may generate a gift message in polite language when the target is a superior (e.g., parents, teachers, bosses, etc.) and in a frank tone when the target is a friend, taking into consideration the relationship between the user and the target. The control unit 230 may also generate a gift message using a tone of voice or buzzwords that are often used by the same group, taking into consideration the age and sex of the user. Furthermore, the control unit 230 may analyze the UGC created by the user, identify the habits (tone of voice, frequency of stamp use, etc.) that the user uses when creating text, and generate a gift message that reflects the habits. In this way, the control unit 230 can generate a gift message that sounds like it was created directly by the user.

Furthermore, the gift message creation page 1700 may include an input area 1720 for selecting the message recipient, the mood of the message, etc. As shown in FIG. 17B, when the user selects message recipient information (e.g., a colleague) and mood information (e.g., a humorous message) in the input area, the control unit 230 may generate and provide a message 1732 based on the selected information.

Furthermore, on the gift message creation page 1700, an image 1710 to be included in the gift message may be selected.

On the other hand, although not shown, the control unit 230 may provide the user terminal with a message list of multiple sample gift messages corresponding to various targets and situations. The control unit 230 may identify the gift message by selecting at least one of the message lists on the user terminal.

As described above, the product recommendation method and system according to the present invention can improve the convenience of gift purchasing for users by providing services ranging from product recommendation, purchase, and gift message generation based on user story information.

Furthermore, the product recommendation method and system according to the present invention can extract recommended products using a large language model that takes as input a prompt containing story information.

Here, the story information is configured to include information about the recipient of the recommended product and the intention of the purchaser to purchase the recommended product, so that the present invention can recommend a product that reflects the intention of the user to purchase the recommended product.

In this way, the product recommendation method and system according to the present invention can recommend products based on the story of the user who will give the recommended product, from a large language model, by constructing a prompt using story information that reflects the recipient of the recommended product and the intention of giving the recommended product. Therefore, according to the present invention, it is possible to reduce the time it takes to select a product to give as a gift, and to purchase a product that is optimized for the recipient and the intention of the gift.

On the other hand, the present invention described above may be realized as a program that can be executed by one or more processes on a computer and stored on a computer-readable medium (or recording medium).

Furthermore, the present invention described above may be realized as computer-readable code or commands on a medium on which a program is recorded. In other words, the present invention may be provided in the form of a program.

Computer-readable media includes any type of recording device on which data that can be read by a computer system is stored. Examples of computer-readable media include hard disk drives (HDDs), solid state disks (SSDs), silicon disk drives (SDDs), ROMs, RAMs, CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices.

Furthermore, the computer-readable medium includes storage, including a server or cloud storage that the electronic device can access via communication. In this case, the computer can download the program according to the present invention from the server or cloud storage via wired or wireless communication.

Furthermore, in the present invention, the aforementioned computer is an electronic device equipped with a processor, i.e., a CPU (Central Processing Unit), and the type of the processor may be any.

The above detailed description should not be interpreted as restrictive in any respect, but should be considered as illustrative. The scope of the present invention should be determined by a reasonable interpretation of the claims, and all modifications within the scope of the equivalents of the present invention are included in the present invention.

10 Large language model 200 Product recommendation system 220 Storage unit 230 Control unit 411, 412, 413 Sample data set 414, 1111, 1220, 1650 Target data set 410, 1010, 1040, 1110, 1210, 1640 Prompt 500, 610, 710, 800, 910, 920, 1310, 1350 Specific page 611, 612, 613, 616, 1732 Message 711 First input area 712 Second input area 913 Sample story information 1014a Text 1032 User information

Claims (16)

A product recommendation method executed by a product recommendation system ,
identifying target story information including at least one of target audience information and target intent information;
The product recommendation system constructs a prompt including at least one sample data set and the target story information;
the product recommendation system processes the prompt as an input to a pre-specified large language model;
obtaining a recommended product category corresponding to the target story information using the large language model with the sample data set and the target story information input by the prompt ;
extracting at least one recommended product that matches the target story information from among the products that belong to the recommended product category;
The sample data set is:
Sample story information including at least one of sample subject information and sample intent information, and a sample product category corresponding to the sample story information;
In the step of acquiring a recommended product category,
A product recommendation method comprising the steps of: referring to the sample data set; and obtaining the recommended product category corresponding to the target story information . The recommended product categories are:
The method of claim 1, wherein the product recommendation method is generated as an output of the large language model. The product recommendation method according to claim 1, characterized in that at least one of the sample story information and the sample product category is obtained as an output of the large language model by inputting a prompt configured based on UGC (User Generated Content) created by multiple users on a page accessible via the Internet into the large language model. The target story information is identified based on a product recommendation request received from a specific user account;
In the step of identifying target story information,
The product recommendation method of claim 1, further comprising identifying the target story information using user input information received from the specific user account and user information stored in a specific server in association with the specific user account. The user input information is
The specific user account is received on a specific page provided to a user terminal logged in,
The specific page is
5. The product recommendation method of claim 4, further comprising: a first input area for receiving input of the target information corresponding to a recipient to whom the recommended product is to be presented; and a second input area for receiving input of the target intention information corresponding to an intention of presenting the recommended product to the recipient. The step of identifying target story information includes:
generating, as the target story information, a sentence including the target subject information and the target intention information received in the first input area and the second input area, using a natural language processing algorithm;
The product recommendation method according to claim 5 , wherein the sentence is included in the prompt. The user information includes at least one of a gender, an age, a residential area, and an occupation of a specific user corresponding to the specific user account;
In the step of generating the sentence,
The product recommendation method according to claim 6 , wherein the sentence is generated so that the user information is included in the target story information. In the step of extracting the recommended products,
The product recommendation method according to claim 1, characterized in that at least one of the multiple products belonging to the recommended product category is extracted as the recommended product in order of highest similarity based on the similarity between the target story information and the multiple products belonging to the recommended product category. The plurality of products are configured to be available for purchase at at least one product sales site accessible via the Internet;
The method of claim 8, further comprising the step of providing, when the recommended products are extracted, product sales information linked to sales pages of the recommended products to a user terminal. The product recommendation method according to claim 9, further comprising a step of generating a message regarding the recommended product using the target story information and information regarding the recommended product when the recommended product is purchased on the sales page. In the step of generating the message,
The product recommendation method of claim 10, wherein the message is generated such that the target information includes information regarding a recipient to whom the recommended product is to be gifted, and the target intention information includes information regarding an intention to gift the recommended product. A product recommendation method executed by a product recommendation system ,
receiving target story information from a user terminal, the target story information including information regarding at least one of a subject of a gift and an intention of the gift;
constructing a first prompt using at least one data set including sample story information and sample product category theme information and the target story information to be input to a pre-specified large language model;
processing the first prompt as an input to the large pre-specified language model;
obtaining a plurality of theme keywords corresponding to the target story information using the large language model with the data set and the target story information input by the first prompt ;
providing the user terminal with the plurality of theme keywords, and selecting at least one of the plurality of theme keywords at the user terminal;
constructing a second prompt at the user terminal that includes the selected theme keyword;
and obtaining a recommended gift category corresponding to the target story information from the large language model by processing the second prompt including the specific theme keyword as an input to the large language model. The second prompt:
at least one first data set including the sample story information, the theme information of the sample product category, and the sample product category;
The method of claim 12, further comprising: a second data set including the target story information and the specific theme keyword. The method of claim 13, further comprising: extracting at least one recommended product corresponding to the target story information from among the products belonging to the recommended gift category. In product recommendation systems,
A storage unit for storing sample story information including at least one of sample subject information and sample intent information, and a sample data set including sample product categories corresponding to the sample story information;
A product recommendation system comprising: a process for identifying target story information including at least one of target subject information and target intention information; a process for constructing a prompt including at least one sample dataset and the target story information; a process for processing the prompt as an input to a pre-specified large language model; a process for obtaining a recommended product category corresponding to the target story information by referring to the sample story information and sample product categories corresponding to the sample story information using the large language model to which the prompt has been input; and a control unit for performing a process for extracting at least one recommended product that matches the target story information from among products belonging to the recommended product category. A program executed by one or more processes in an electronic device,
identifying target story information including at least one of target audience information and target intent information;
constructing at least one sample data set and a prompt including the target story information;
processing the prompt as input to a pre-specified large language model;
obtaining a recommended product category corresponding to the target story information using the large language model with the sample data set and the target story information input by the prompt ;
and extracting at least one recommended product that matches the target story information from among the products that belong to the recommended product category .
The sample data set is:
Sample story information including at least one of sample subject information and sample intent information, and a sample product category corresponding to the sample story information;
In the step of acquiring a recommended product category,
A program characterized by referring to the sample data set and obtaining the recommended product category corresponding to the target story information .