JP7664934B2 - Recommendation device - Google Patents

Description

This disclosure relates to a recommendation device.

There is known a technology that recommends appropriate financial products such as insurance to a user. For example, Patent Document 1 describes an insurance proposal server that calculates risk based on user information, estimates insurance combinations by inputting the risk into a proposal model, and presents the insurance combinations to the user.

JP 2020-71622 A

Even if the risk is not that high, such as in the case of a disaster, if there is a high possibility of suffering a large loss, it is conceivable to take out insurance to compensate for the loss. However, in the insurance proposal server described in Patent Document 1, risk is used as an input for the proposal model, so insurance that is correlated with risk is proposed. Therefore, there is a risk that appropriate insurance will not be proposed to the user.

This disclosure describes a recommendation device that can recommend more appropriate financial products.

A recommendation device according to one aspect of the present disclosure is a device that recommends financial products to a user from among a plurality of financial products. This recommendation device includes a determination unit that determines a recommendation target, which is a financial product to be recommended to a user from among the plurality of financial products, based on a purchase score and a purchase amount set for each of the plurality of financial products, which are indicative of the likelihood that the user will purchase the financial product, and a purchase amount suitable for the user, and an output unit that outputs recommendation information regarding the recommendation target.

In this recommendation device, a recommended item is determined from among a plurality of financial products based on the purchase score and purchase amount, and recommendation information is output. A financial product with a high purchase score is more likely to be purchased by a user. On the other hand, even if the purchase score is low, a high return can be expected if the purchase amount of the financial product is high. Therefore, by determining a recommended item taking into account not only the purchase score but also the purchase amount, it is possible to recommend more appropriate financial products to the user.

This disclosure makes it possible to recommend more appropriate financial products.

FIG. 1 is a schematic configuration diagram of a recommendation system including a recommendation server according to an embodiment. Fig. 2(a) is a diagram showing an example of user basic information stored in the user information DB (database) shown in Fig. 1. Fig. 2(b) is a diagram showing an example of location information stored in the user information DB shown in Fig. 1. Fig. 2(c) is a diagram showing an example of payment information stored in the user information DB shown in Fig. 1. FIG. 3 is a diagram showing an example of insurance information stored in the insurance information DB shown in FIG. FIG. 4 is a block diagram showing the functional configuration of the recommendation server shown in FIG. FIG. 5 is a sequence diagram showing a series of processes of the recommendation method performed by the recommendation system shown in FIG. FIG. 6 is a flowchart showing in detail the determination process shown in FIG. FIG. 7 is a diagram for explaining the process of determining a recommendation target. FIG. 8 is a diagram showing an example of a display screen for recommendation information. FIG. 9 is a diagram for explaining a modified example of the process of determining a recommendation target. FIG. 10 is a diagram illustrating a hardware configuration of the recommendation server illustrated in FIG.

Hereinafter, an embodiment of the present disclosure will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements are given the same reference numerals and duplicated descriptions are omitted.

With reference to Figures 1 to 3, the configuration of a recommendation system including a recommendation device according to one embodiment will be described. Figure 1 is a schematic configuration diagram of a recommendation system including a recommendation device according to one embodiment. (a) of Figure 2 is a diagram showing an example of basic user information stored in a user information DB (database) shown in Figure 1. (b) of Figure 2 is a diagram showing an example of location information stored in the user information DB shown in Figure 1. (c) of Figure 2 is a diagram showing an example of payment information stored in the user information DB shown in Figure 1. Figure 3 is a diagram showing an example of insurance enrollment information stored in the insurance enrollment information DB shown in Figure 1.

The recommendation system 1 shown in FIG. 1 is a system for recommending financial products to a user. In this embodiment, an insurance product will be used as the financial product.

The recommendation system 1 includes a plurality of terminal devices 2, a user information DB 3, an insurance information DB 4, and a recommendation device 10. The plurality of terminal devices 2, the user information DB 3, the insurance information DB 4, and the recommendation device 10 are configured to be able to communicate with each other via a network NW. The network NW may be configured as either a wired or wireless network. Examples of the network NW include a mobile communication network, the Internet, and a WAN (Wide Area Network). The following explanation focuses mainly on one terminal device 2, but the same applies to the other terminal devices 2.

The terminal device 2 is a device used by a user. Examples of the terminal device 2 include a smartphone, a tablet terminal, a notebook PC (Personal Computer), and a desktop PC.

The terminal device 2 acquires location information (latitude and longitude) of the terminal device 2 using a GPS (Global Positioning System) or the like. The terminal device 2 may acquire information on the installation location of a master station of a connected wireless network as location information. Examples of the installation location of a master station include a mobile network base station and a Wi-Fi access point. The terminal device 2 may acquire location information of a terminal present in the vicinity of the terminal device 2 as location information of the terminal device 2. Examples of such terminals include a Bluetooth (registered trademark) beacon terminal. Details of the location information will be described later. The terminal device 2 periodically transmits the location information to the user information DB 3.

The terminal device 2 generates payment information regarding a payment made by a user using the terminal device 2. For example, when a user purchases a product using a payment application installed on the terminal device 2, the terminal device 2 generates the payment information. Details of the payment information will be described later. For example, the terminal device 2 transmits the payment information to the user information DB 3 each time it generates payment information.

User information DB3 is a database that stores user information for each user. User information is information about a user, and includes user basic information, location information, and payment information. User information may further include other information, such as the usage history (log) of the terminal device 2. User basic information is basic information about a user. As shown in (a) of FIG. 2, the user basic information includes a user ID (identifier), a terminal ID, gender, and age. The user ID is information that can uniquely identify a user. The terminal ID is information that can uniquely identify a terminal device 2. Here, the terminal ID indicates the terminal device 2 used by the user identified by the user ID. The user basic information may further include other information. The user basic information is, for example, set in advance by the user.

Location information is information that indicates the location of each terminal device. As shown in (b) of FIG. 2, location information includes a terminal ID, the time (timestamp) when the location information was acquired, latitude, and longitude. When user information DB3 receives location information from each terminal device 2, it stores the received location information. User information DB3 stores multiple pieces of location information for each terminal device 2 as a history (log) of location information.

The payment information is information relating to payments made using each terminal device 2. As shown in (c) of FIG. 2, the payment information includes the terminal ID, the time the payment was made, the location where the payment was made, the amount, and the product name. When the user information DB3 receives payment information from each terminal device 2, it stores the received payment information. The user information DB3 stores multiple payment information for each terminal device 2 as a history of payment information.

Insurance information DB4 is a database that stores insurance information for each user. The insurance information is information regarding the insurance products to which each user has subscribed. As shown in FIG. 3, the insurance information includes an insurance ID, a user ID, and an insurance premium. The insurance ID is information that can uniquely identify an insurance product. The insurance premium is the amount that a user identified by the user ID is paying for the insurance product identified by the insurance ID. The insurance premium is, for example, a monthly insurance premium. The insurance information may include the number of shares purchased instead of the insurance premium, or may include the number of shares purchased together with the insurance premium.

The recommendation device 10 is a device that recommends insurance products (recommendation targets) to a user from among multiple insurance products. An example of the recommendation device 10 is an information processing device such as a server device.

The functional configuration of the recommendation device 10 will be described with reference to FIG. 4. FIG. 4 is a block diagram showing the functional configuration of the recommendation device shown in FIG. 1. As shown in FIG. 4, the recommendation device 10 functionally includes an acquisition unit 11, a generation unit 12, a calculation unit 13 (first calculation unit), an enrollment score memory unit 14, a calculation unit 15 (second calculation unit), an insurance premium memory unit 16, a reception unit 17, a determination unit 18, and an output unit 19.

The acquisition unit 11 is a functional unit that acquires user information and insurance information. The acquisition unit 11 acquires user information from the user information DB3 and acquires insurance information from the insurance information DB4.

The generation unit 12 is a functional unit that generates an enrollment prediction model and an insurance premium prediction model. The enrollment prediction model is a machine learning model with feature quantities generated from user information as explanatory variables and an insurance product enrollment score as a target variable, and is configured, for example, by a neural network. The enrollment score is a value that indicates the likelihood that a user will enroll in an insurance product. The enrollment score is, for example, a numerical value in the range of 0 to 1. For example, the higher the enrollment score of an insurance product, the more likely the user will enroll in that insurance product. The generation unit 12 generates an enrollment prediction model for each insurance product by performing machine learning for each insurance product.

The insurance premium prediction model is a machine learning model that uses features generated from user information as explanatory variables and the insurance premium paid for an insurance product as the objective variable, and is configured, for example, by a neural network. The insurance premium is the premium paid by a user for an insurance product, and is obtained, for example, by multiplying the insurance premium per policy by the number of policies purchased. The generation unit 12 generates an insurance premium prediction model for each insurance product by performing machine learning for each insurance product. The method of generating features, the method of generating an enrollment prediction model, and the method of generating an insurance premium prediction model will be described later.

The calculation unit 13 is a functional unit that calculates an enrollment score for each of a plurality of insurance products based on user information. The calculation unit 13 calculates the enrollment score using an enrollment prediction model. The calculation unit 13 generates features from the user information and obtains the enrollment score from the enrollment prediction model by inputting the generated features into the enrollment prediction model.

The enrollment score memory unit 14 is a functional unit that stores the enrollment score for each insurance product of each user. The enrollment score memory unit 14 stores, for example, a data set that associates a user ID, an insurance ID, and an enrollment score.

The calculation unit 15 is a functional unit that calculates insurance premiums appropriate for a user for each of multiple insurance products based on user information. The calculation unit 15 calculates insurance premiums using an insurance premium prediction model. The calculation unit 15 generates features from the user information and obtains insurance premiums from the insurance premium prediction model by inputting the generated features into the insurance premium prediction model.

The insurance premium memory unit 16 is a functional unit that stores the insurance premiums for each insurance product of each user. The insurance premium memory unit 16 stores, for example, a data set that associates a user ID, an insurance ID, and an insurance premium.

The receiving unit 17 is a functional unit that receives a recommendation request from the terminal device 2. The recommendation request is a command to request recommendation information for insurance products. The recommendation request includes the user ID of the user requesting the recommendation information.

The determination unit 18 is a functional unit that determines a recommendation target, which is an insurance product to be recommended to a user, from among a plurality of insurance products. The determination unit 18 determines a recommendation target based on the enrollment score and insurance premium set for each of the plurality of insurance products. The determination unit 18 compares the enrollment score with an enrollment score threshold Rth (first threshold), and compares the insurance premium with an insurance premium threshold Cth (second threshold), and determines a recommendation target according to the comparison result. The thresholds Rth and Cth are set in advance. The determination unit 18 determines, for example, an insurance product that satisfies at least one of the conditions that the enrollment score is greater than the threshold Rth (first condition) and that the insurance premium is greater than the threshold Cth (second condition) as a recommendation target.

The output unit 19 is a functional unit that outputs recommendation information related to the recommendation target. The output unit 19, for example, outputs (transmits) the recommendation information to the terminal device 2. The output unit 19 may output the recommendation information to a memory (not shown) within the recommendation device 10.

Next, the recommendation method performed by the recommendation system 1 (recommendation device 10) will be described with reference to Figures 5 to 8. Figure 5 is a sequence diagram showing a series of processes of the recommendation method performed by the recommendation system shown in Figure 1. Figure 6 is a flowchart showing in detail the decision process shown in Figure 5. Figure 7 is a diagram for explaining the decision process of the recommendation target. Figure 8 is a diagram showing an example of a display screen for recommendation information.

As shown in FIG. 5, first, the acquisition unit 11 of the recommendation device 10 sends an acquisition request for user information to the user information DB 3 (step S1). In step S1, the acquisition unit 11 may send an acquisition request to acquire user information of all users, or may send an acquisition request to acquire user information of some users. Then, when the user information DB 3 receives the acquisition request for user information from the recommendation device 10, it sends the requested user information to the recommendation device 10 (step S2).

Next, the acquisition unit 11 of the recommendation device 10 transmits an acquisition request for insurance information to the insurance information DB 4 (step S3). In step S3, the acquisition unit 11 transmits an acquisition request to acquire insurance information for all insurance products that can be recommended, for example. Then, when the insurance information DB 4 receives the acquisition request for insurance information from the recommendation device 10, it transmits the requested insurance information to the recommendation device 10 (step S4).

Then, the acquisition unit 11 of the recommendation device 10 receives user information from the user information DB 3, and when it receives insurance information from the insurance information DB 4, it outputs the user information and the insurance information to the generation unit 12. Next, when the generation unit 12 receives the user information and the insurance information from the acquisition unit 11, it generates an enrollment prediction model (step S5). In step S5, the generation unit 12 generates an enrollment prediction model for each insurance product by performing machine learning for each insurance product. The machine learning is performed, for example, using a GBDT (Gradient Boosting Decision Tree) algorithm. For example, in the machine learning, a set of a feature generated from the user information of a user who has previously enrolled in an insurance product and an enrollment score (= 1) of the insurance product is used as correct answer data, and a set of a feature generated from the user information of a user who has not enrolled in an insurance product and an enrollment score (= 0) of the insurance product is used as incorrect answer data. Then, the generation unit 12 outputs the enrollment prediction model to the calculation unit 13.

Here, an example of a method for generating features will be described. The generation unit 12 uses gender and age from the user information as features. The generation unit 12 may estimate the places where the user stayed and the duration of stay from the time-series location information of the terminal device 2, and use the places and duration of stay as features. Furthermore, in order to reduce the impact on the subscription score of places that the user does not usually visit but happens to stay in, temporal changes in the places and duration of stay may be used as features. From the payment information of the terminal device 2, the generation unit 12 calculates the number of payments, the number of stores where payments were made, and the total payment amount as features. The amount for each category of paid products (services) may be used as a feature.

Furthermore, the generation unit 12 generates an insurance premium prediction model (step S6). In step S6, the generation unit 12 generates an insurance premium prediction model for each insurance product by performing machine learning for each insurance product. The machine learning is performed using, for example, the GBDT algorithm. In the machine learning, for example, a pair of features generated from user information of a user who has previously subscribed to an insurance product and the insurance premium paid by that user for the insurance product is used as correct answer data. The method of generating the features is as described above. Then, the generation unit 12 outputs the insurance premium prediction model to the calculation unit 15.

Next, the acquisition unit 11 transmits an acquisition request to acquire user information of all users to the user information DB 3 (step S7). Then, when the user information DB 3 receives the acquisition request for user information from the recommendation device 10, it transmits the requested user information to the recommendation device 10 (step S8). Then, when the acquisition unit 11 receives the user information from the user information DB 3, it outputs the user information to the calculation unit 13 and the calculation unit 15.

Next, when the calculation unit 13 receives the user information from the acquisition unit 11, it calculates the subscription score of each user for each of the multiple insurance products (step S9). In step S9, the calculation unit 13 calculates the subscription score using the subscription prediction model. Specifically, the calculation unit 13 generates features from the user information of each user in the same manner as the generation unit 12 generates features. Then, the calculation unit 13 inputs the features for each user into the subscription prediction model of each insurance product and obtains a subscription score output from each subscription prediction model. Then, the calculation unit 13 outputs a data set in which the user ID, insurance ID, and subscription score are associated with each other to the subscription score storage unit 14, and stores the data in the subscription score storage unit 14.

Next, when the calculation unit 15 receives the user information from the acquisition unit 11, it calculates the insurance premium suitable for each user for each of the multiple insurance products (step S10). In step S10, the calculation unit 15 calculates the insurance premium using an insurance premium prediction model. Specifically, the calculation unit 15 generates features from the user information of each user in the same manner as the generation unit 12 generates features. Then, the calculation unit 15 inputs the features for each user into the insurance premium prediction model of each insurance product and obtains the insurance premium output from each insurance premium prediction model. Then, the calculation unit 15 outputs a data set in which the user ID, insurance ID, and insurance premium are associated to the insurance premium memory unit 16, and stores the data in the insurance premium memory unit 16.

Next, the terminal device 2 transmits a recommendation request to the recommendation device 10 (step S11). Then, when the receiving unit 17 of the recommendation device 10 receives the recommendation request transmitted from the terminal device 2, it outputs the user ID included in the recommendation request to the determining unit 18.

Next, when the determination unit 18 receives the user ID from the receiving unit 17, it performs a determination process (step S12). As shown in FIG. 6, in the determination process of step S12, the determination unit 18 first obtains the enrollment score for each insurance product of the user identified by the user ID (step S21). Specifically, the determination unit 18 obtains a set of the insurance ID and the enrollment score associated with the user ID received from the receiving unit 17 from the enrollment score storage unit 14. Then, the determination unit 18 obtains the insurance premium for each insurance product of the user identified by the user ID (step S22). Specifically, the determination unit 18 obtains a set of the insurance ID and the insurance premium associated with the user ID received from the receiving unit 17 from the insurance premium storage unit 16.

Next, the determination unit 18 determines a recommendation target from among the multiple insurance products (step S23). In step S23, the determination unit 18 determines a recommendation target based on the enrollment score and insurance premium set for each insurance product using the set of insurance ID and enrollment score acquired from the enrollment score memory unit 14 and the set of insurance ID and insurance premium acquired from the insurance premium memory unit 16. Specifically, the determination unit 18 compares the enrollment score of each insurance product with a threshold value Rth and compares the insurance premium of each insurance product with a threshold value Cth, and determines a recommendation target according to the comparison results.

For example, as shown in FIG. 7, the determination unit 18 determines, as a recommendation target, an insurance product that satisfies at least one of the conditions that the enrollment score is greater than a threshold value Rth and that the insurance premium is greater than a threshold value Cth. In other words, if areas R1, R2, and R4 contain points indicating the enrollment score and insurance premium of an insurance product, the insurance product is determined as a recommendation target, and if area R3 contains points indicating the enrollment score and insurance premium of an insurance product, the insurance product is not a recommendation target.

Region R1 is a region where the enrollment score is greater than the threshold Rth and the insurance premium is less than the threshold Cth (less than or equal to the threshold Cth). Region R2 is a region where the enrollment score is greater than the threshold Rth and the insurance premium is greater than the threshold Cth. Region R3 is a region where the enrollment score is less than the threshold Rth (less than or equal to the threshold Rth) and the insurance premium is less than the threshold Cth (less than or equal to the threshold Cth). Region R4 is a region where the enrollment score is less than the threshold Rth (less than or equal to the threshold Rth) and the insurance premium is greater than the threshold Cth.

In the example shown in Figure 7, point Ia indicating the enrollment score and insurance premium of insurance product A is included in area R1. Points Ib and If indicating the enrollment score and insurance premium of insurance products B and F are included in area R2. Points Ic and Ie indicating the enrollment score and insurance premium of insurance products C and E are included in area R3. Point Id indicating the enrollment score and insurance premium of insurance product D is included in area R4. Therefore, insurance products A, B, D, and F are determined to be recommended targets.

Furthermore, when multiple insurance products are determined as the recommended products, the determination unit 18 determines the priority order of the recommended products. For example, the determination unit 18 determines the priority order of the recommended products in the following order: an insurance product that satisfies both the condition that the subscription score is greater than a threshold value Rth (first condition) and the condition that the insurance premium is greater than a threshold value Cth (second condition), an insurance product that satisfies the first condition but does not satisfy the second condition, and an insurance product that does not satisfy the first condition but satisfies the second condition.

Next, the determination unit 18 generates recommendation information regarding the recommended target (step S24). For example, the determination unit 18 generates recommendation information including the name of the recommended insurance product and the insurance premium. The recommendation information may further include a graph showing the premium and risk of each insurance product. Note that the enrollment score of an insurance product can be considered as a risk to the subject of coverage of the insurance product, so the determination unit 18 creates a graph using the enrollment score of each insurance product as a risk. Then, the determination unit 18 outputs the recommendation information to the output unit 19.

Next, the output unit 19 transmits the recommendation information to the terminal device 2 (step S13). Then, when the terminal device 2 receives the recommendation information transmitted from the recommendation device 10, it displays the recommendation information on the display. For example, as shown in FIG. 8, the names and premiums of the recommended insurance products are displayed together with a graph showing the premiums and risks of each insurance product. The names and premiums of the recommended insurance products are arranged from top to bottom according to the priority order of the recommended products.

This completes the series of processes in the recommendation method. Steps S1 to S10 are performed in advance before receiving a recommendation request from the terminal device 2 (offline processing). Steps S3 and S4 may be performed before steps S1 and S2, or in parallel with steps S1 and S2. Step S6 may be performed before step S5, or in parallel with step S5. Step S10 may be performed before step S9, or in parallel with step S9. Step S22 may be performed before step S21, or in parallel with step S21.

After step S11, steps S7 to S10 may be performed. In this case, in step S7, the acquisition unit 11 sends an acquisition request to the user information DB 3 to acquire user information of the user identified by the user ID included in the recommendation request, and in step S8, the user information DB 3 sends the user information of the requested user to the recommendation device 10. Furthermore, in step S9, the calculation unit 13 calculates the subscription score of the user identified by the user ID included in the recommendation request, and outputs the subscription score to the determination unit 18. In step S10, the calculation unit 15 calculates an insurance premium suitable for the user identified by the user ID included in the recommendation request, and outputs the insurance premium to the determination unit 18.

In the recommendation device 10 described above, a recommended target is determined from among multiple insurance products based on the enrollment score and insurance premium, and recommendation information is output. A user is more likely to enroll in an insurance product with a high enrollment score. On the other hand, even if the enrollment score is low, a high return can be expected if the insurance premium for the insurance product is high. If the insurance premium for an insurance product suitable for a user is high, it can be said that the frequency with which losses compensated by that insurance product occur to the user or the amount of loss suffered by the user is high. Therefore, by determining a recommended target taking into account not only the enrollment score but also the insurance premium, it is possible to recommend a more appropriate insurance product to the user.

The determination unit 18 compares the subscription score with a threshold value Rth and the insurance premium with a threshold value Cth, and determines the recommended target according to the comparison results. Specifically, the determination unit 18 determines, as the recommended target, an insurance product that satisfies at least one of the first condition that the subscription score is greater than the threshold value Rth and the second condition that the insurance premium is greater than the threshold value Cth. As described above, the user is more likely to subscribe to an insurance product with a high subscription score. When the insurance premium suitable for the user is high, it can be said that the frequency with which the loss compensated by the insurance product occurs to the user or the amount of loss suffered by the user is large. Therefore, by determining, as the recommended target, an insurance product that satisfies at least one of the first condition and the second condition, it is possible to recommend a more appropriate insurance product to the user.

The determination unit 18 determines the priority order of the recommended items so that the order is insurance items that satisfy both the first condition and the second condition, insurance items that satisfy the first condition but do not satisfy the second condition, and insurance items that do not satisfy the first condition but do satisfy the second condition. An insurance item with a high subscription score and a high premium is considered to be the most important (essential) insurance item for the user. An insurance item with a high subscription score and a low premium is not essential, but is recommended for subscription because a user who has something in common with the user has subscribed to the insurance item. An insurance item with a low subscription score and a high premium does not need to be subscribed to because a user who has something in common with the user has not subscribed to the insurance item, but it is considered to mean that the frequency of occurrence of losses or the amount of losses compensated by the insurance item is large, so it is safe to subscribe to the insurance item. Therefore, according to the above configuration, the recommended items can be ranked according to their importance to the user, so it is possible to recommend a more suitable insurance item to the user.

The calculation unit 13 calculates an enrollment score for each of a plurality of insurance products based on the user information. Users who share commonalities in gender, age, behavior, etc. are considered to be equally likely to enroll in insurance products. Therefore, by using the user information, the enrollment score for each insurance product can be calculated with high accuracy.

The calculation unit 15 calculates insurance premiums appropriate for the user for each of multiple insurance products based on the user information. Users who share commonalities in gender, age, behavior, etc. are expected to pay similar amounts of insurance premium for insurance products. Therefore, by using the user information, the insurance premium for each insurance product can be calculated with high accuracy.

Although the above describes embodiments of the present disclosure, the present disclosure is not limited to the above embodiments.

The recommendation device 10 may be configured by one device that is physically or logically coupled, or may be configured by multiple devices that are physically or logically separated from each other. For example, the recommendation device 10 may be realized by multiple computers distributed on a network, such as in cloud computing. As described above, the configuration of the recommendation device 10 may include any configuration that can realize the functions of the recommendation device 10.

In the above embodiment, an insurance product is used as the financial product, but the recommendation device 10 may determine the recommendation target from other financial products such as stocks, bonds, and investment trusts. When the recommendation target is generalized to a financial product, the subscription score is generalized to a purchase score indicating the possibility that the user will purchase the financial product, and the insurance premium is the purchase amount (number of purchases) paid for the financial product, and is generalized to a purchase amount (number of purchases) appropriate for the user.

The recommendation device 10 may not include the calculation unit 13 and the enrollment score memory unit 14. In this case, the determination unit 18 may acquire a set of insurance ID and enrollment score associated with the user ID included in the recommendation request from an external enrollment score memory unit. The recommendation device 10 may not include the calculation unit 15 and the insurance premium memory unit 16. In this case, the determination unit 18 may acquire a set of insurance ID and insurance premium associated with the user ID included in the recommendation request from an external insurance premium memory unit.

The recommendation device 10 may not be equipped with the generation unit 12. In this case, the calculation unit 13 calculates the enrollment score using a pre-generated enrollment prediction model. Similarly, the calculation unit 15 calculates the insurance premium using a pre-generated insurance premium prediction model. The calculation unit 13 may calculate the enrollment score on a rule-based basis based on user information without using the enrollment prediction model. The calculation unit 15 may calculate the insurance premium on a rule-based basis based on user information without using the insurance premium prediction model.

The recommendation device 10 may not include the acquisition unit 11, the generation unit 12, the calculation unit 13, the enrollment score memory unit 14, the calculation unit 15, and the insurance premium memory unit 16. In this case, the determination unit 18 may acquire a set of an insurance ID and an enrollment score associated with a user ID included in the recommendation request from an external enrollment score memory unit, and may acquire a set of an insurance ID and an insurance premium associated with a user ID included in the recommendation request from an external insurance premium memory unit.

The determination unit 18 may calculate a value (Sharpe ratio) obtained by dividing the insurance premium by the enrollment score for each insurance product, and determine the recommended target based on the value. The enrollment score can be seen as the risk of occurrence of a loss compensated by the insurance product. Therefore, the larger the value, the higher the degree of risk compensation. The determination unit 18 determines, for example, an insurance product whose value is greater than a predetermined threshold as a recommended target. According to this configuration, a new index (Sharpe ratio) is generated using the prediction result (enrollment score) of the enrollment prediction model and the prediction result (premium) of the premium prediction model. Therefore, the recommended target can be determined simply by comparing one value with the threshold, which simplifies the process of determining the recommended target.

Furthermore, when multiple insurance products are determined to be recommended, the determination unit 18 may give a higher priority to the insurance product the smaller the above value. In this case, insurance products with low insurance premiums and high risks are preferentially recommended. This can therefore encourage users to subscribe to insurance products. As described above, the priority can be determined based on only the magnitude of a single value, making it possible to simplify the priority determination process.

The determination unit 18 may obtain a regression line showing the distribution trend of the enrollment scores and insurance premiums based on the enrollment scores and insurance premiums for multiple insurance products, and determine the recommendation target based on the distance between the enrollment score and insurance premium of each of the multiple insurance products and the regression line. A specific explanation will be given with reference to Figure 9. Figure 9 is a diagram for explaining a modified example of the process of determining the recommendation target.

The determination unit 18 obtains a regression line RL from points Ig to Ij indicating the enrollment scores and insurance premiums of insurance products G to J. The determination unit 18 then calculates the distance between each of the points Ig to Ij and the regression line RL, and compares each distance with a distance threshold. The determination unit 18 then determines, as a recommendation target, an insurance product corresponding to a point whose distance is shorter (smaller) than the threshold. In the example shown in FIG. 9, the distance between point Ih and the regression line RL is the shortest, and the distances between point Ii, point Ig, and point Ij are larger in this order. It is assumed that only the distance between point Ih and the regression line RL and the distance between point Ii and the regression line RL are smaller than the threshold. In this case, the determination unit 18 determines insurance product H and insurance product I as recommendation targets. Furthermore, the determination unit 18 may give a higher priority to an insurance product as the distance between the point indicating the enrollment score and insurance premium of the insurance product and the regression line RL is shorter. Therefore, the determination unit 18 generates recommendation information such that the priority order is given to insurance product H and insurance product I.

It is believed that the enrollment score and insurance premiums of an insurance product are positively correlated. Therefore, the regression line RL is believed to indicate the relationship between the enrollment score and insurance premiums that is optimal for the user. In other words, the closer the point indicating the enrollment score and insurance premium is to the regression line RL, the better the balance between the enrollment score and insurance premiums is for the user. Therefore, by determining the recommended target according to the distance between the point indicating the enrollment score and insurance premium and the regression line RL, it is possible to recommend a recommended target that is suitable for the user.

The block diagrams used to explain the above embodiments show functional blocks. These functional blocks (components) are realized by any combination of at least one of hardware and software. The method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or may be realized using two or more devices that are physically or logically separated and directly or indirectly connected (e.g., using wires, wirelessly, etc.). The functional blocks may be realized by combining software with the one device or the multiple devices.

Functions include, but are not limited to, judgment, determination, judgement, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, resolution, selection, election, establishment, comparison, assumption, expectation, regard, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, and assignment. For example, a functional block (component) that performs the transmission function is called a transmitting unit or transmitter. In either case, as described above, there are no particular limitations on the method of implementation.

For example, the recommendation device 10 in one embodiment of the present disclosure may function as a computer that performs the processing of the present disclosure. FIG. 10 is a diagram showing an example of a hardware configuration of the recommendation device 10 according to one embodiment of the present disclosure. The above-mentioned recommendation device 10 may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, and a bus 1007.

In the following description, the term "apparatus" may be interpreted as a circuit, device, unit, etc. The hardware configuration of the recommendation apparatus 10 may be configured to include one or more of the apparatuses shown in the figure, or may be configured to exclude some of the apparatuses.

Each function in the recommendation device 10 is realized by loading a specific software (program) onto hardware such as the processor 1001 and memory 1002, causing the processor 1001 to perform calculations, control communication via the communication device 1004, and control at least one of the reading and writing of data in the memory 1002 and storage 1003.

The processor 1001, for example, operates an operating system to control the entire computer. The processor 1001 may be configured by a central processing unit (CPU) including an interface with peripheral devices, a control device, an arithmetic unit, and a register. For example, each function of the recommendation device 10 described above may be realized by the processor 1001.

The processor 1001 reads out programs (program codes), software modules, data, etc. from at least one of the storage 1003 and the communication device 1004 into the memory 1002, and executes various processes according to these. As the program, a program that causes a computer to execute at least a part of the operations described in the above embodiment is used. For example, each function of the recommendation device 10 may be realized by a control program stored in the memory 1002 and running on the processor 1001. Although the above-mentioned various processes have been described as being executed by one processor 1001, they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may be transmitted from a network via a telecommunications line.

The memory 1002 is a computer-readable recording medium and may be composed of at least one of, for example, a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), and a random access memory (RAM). The memory 1002 may be called a register, a cache, or a main memory (main storage device). The memory 1002 can store executable programs (program codes), software modules, etc. for implementing the recommendation method according to one embodiment of the present disclosure.

Storage 1003 is a computer-readable recording medium, and may be, for example, at least one of an optical disk such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disk, a digital versatile disk, a Blu-ray (registered trademark) disk), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, and a magnetic strip. Storage 1003 may also be referred to as an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, a server, or other suitable medium including at least one of memory 1002 and storage 1003.

The communication device 1004 is hardware (transmitting/receiving device) for communicating between computers via at least one of a wired network and a wireless network, and is also called, for example, a network device, a network controller, a network card, or a communication module. The communication device 1004 may be configured to include a high-frequency switch, a duplexer, a filter, and a frequency synthesizer to realize at least one of Frequency Division Duplex (FDD) and Time Division Duplex (TDD). For example, the above-mentioned acquisition unit 11, receiving unit 17, and output unit 19 may be realized by the communication device 1004.

The input device 1005 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that accepts input from the outside. The output device 1006 is an output device (e.g., a display, a speaker, a LED lamp, etc.) that performs output to the outside. Note that the input device 1005 and the output device 1006 may be integrated into one configuration (e.g., a touch panel).

Each device, such as the processor 1001 and the memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses between each device.

The recommendation device 10 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), and a field programmable gate array (FPGA), and some or all of the functional blocks may be realized by the hardware. For example, the processor 1001 may be implemented using at least one of these pieces of hardware.

Notification of information is not limited to the aspects/embodiments described in this disclosure and may be performed using other methods.

In the processing procedures, sequences, and flowcharts of each aspect/embodiment described in this disclosure, the order of processing may be changed unless inconsistent. For example, the methods described in this disclosure present elements of various steps using an example order and are not limited to the particular order presented.

Information, etc. may be output from a higher layer to a lower layer, or from a lower layer to a higher layer. Information, etc. may be input/output via multiple network nodes.

The input and output information may be stored in a specific location (e.g., memory) or may be managed using a management table. The input and output information may be overwritten, updated, or added to. The output information may be deleted. The input information may be transmitted to another device.

The determination may be based on a value represented by a single bit (0 or 1), a Boolean (true or false) value, or a numerical comparison (e.g., with a predetermined value).

Each aspect/embodiment described in this disclosure may be used alone, in combination, or switched as the implementation progresses. Notification of specific information (e.g., notification that "X is the case") is not limited to explicit notification, and may be performed implicitly (e.g., by not notifying the specific information).

Although the present disclosure has been described in detail above, it is clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is intended as an illustrative example and does not have any limiting meaning on the present disclosure.

Software shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

Software, instructions, information, and the like may be transmitted and received via a transmission medium. For example, if the software is transmitted from a website, server, or other remote source using wired technologies (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL)), and/or wireless technologies (such as infrared, microwave), these wired and/or wireless technologies are included within the definition of a transmission medium.

The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

In addition, terms explained in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meaning.

As used in this disclosure, the terms "system" and "network" are used interchangeably.

The information, parameters, etc. described in this disclosure may be expressed using absolute values, relative values from a predetermined value, or other corresponding information.

The names used for the parameters described above are not limiting in any way. Furthermore, the formulas etc. using these parameters may differ from the formulas etc. explicitly disclosed in this disclosure.

The terms "determining" and "determining" as used in this disclosure may encompass a wide variety of actions. Each of "determining" and "determining" may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, search, inquiry (e.g., searching in a table, database, or another data structure), and ascertaining. Each of "determining" and "determining" may include receiving (e.g., receiving information), transmitting (e.g., transmitting information), input, output, and accessing (e.g., accessing data in a memory). Each of "determining" and "determining" may include resolving, selecting, choosing, establishing, and comparing. That is, "determining" may include any action that may be considered a "judgment." A "determination" may include any action that may be considered a "determination." "Judgment" may be read as "assuming," "expecting," or "considering," among others.

The terms "connected" and "coupled", or any variation thereof, refer to any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connect" may be read as "access". When "connected" or "coupled" is used in this disclosure, two elements may be considered to be "connected" or "coupled" to each other using at least one of one or more wires, cables, and printed electrical connections, and may be considered to be "connected" or "coupled" to each other using electromagnetic energy having wavelengths in the radio frequency range, microwave range, and light (both visible and invisible) range, as some non-limiting and non-exhaustive examples.

As used in this disclosure, the phrase "based on" does not mean "based only on," unless expressly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."

Any reference to elements using designations such as "first" and "second" used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Thus, a reference to a first and a second element does not imply either that only two elements may be employed or that the first element must precede the second element in some way.

The "parts" in the configuration of each of the above devices may be replaced with "circuits" or "devices", etc.

When used in this disclosure, the terms "include," "including," and variations thereof are intended to be inclusive, similar to the term "comprising." Additionally, the term "or," as used in this disclosure, is not intended to be an exclusive or.

In this disclosure, where articles have been added by translation, such as "a," "an," and "the" in English, this disclosure may include that the nouns following these articles are plural.

In this disclosure, the term "A and B are different" may mean "A and B are different from each other." In addition, the term may mean "A and B are each different from C." Terms such as "separate" and "combined" may also be interpreted in the same way as "different."

1... recommendation system, 2... terminal device, 3... user information DB, 4... insurance enrollment information DB, 10... recommendation device, 11... acquisition unit, 12... generation unit, 13... calculation unit (first calculation unit), 15... calculation unit (second calculation unit), 17... receiving unit, 18... determination unit, 19... output unit, 1001... processor, 1002... memory, 1003... storage, 1004... communication device, 1005... input device, 1006... output device, 1007... bus, Rth... threshold (first threshold), Cth... threshold (second threshold).

Claims (2)

A recommendation device that recommends financial products to a user from among a plurality of financial products,
a first calculation unit that calculates a purchase score indicating a likelihood that the user will purchase the financial product for each of the plurality of financial products based on user information related to the user;
a second calculation unit that calculates a purchase amount appropriate for the user for each of the plurality of financial products based on the user information;
A determination unit that determines a recommendation target, which is a financial product to be recommended to the user, from among the plurality of financial products based on the purchase score and the purchase amount calculated for each of the plurality of financial products;
an output unit that outputs recommendation information related to the recommendation target;
Equipped with
the first calculation unit is a first prediction model provided for each of the plurality of financial products, the first prediction model being a machine learning model in which a feature amount generated from user information of the user is used as an explanatory variable and a purchase score of the financial product is used as a target variable, and the first prediction model inputs a feature amount generated from the user information of the user into the first prediction model, thereby obtaining the purchase score from the first prediction model;
the second calculation unit is a second prediction model provided for each of the plurality of financial products, the second prediction model being a machine learning model in which a feature amount generated from user information of the user is used as an explanatory variable and a purchase amount is used as a target variable, thereby obtaining the purchase amount from the second prediction model;
The determination unit compares the purchase score with a first threshold and the purchase amount with a second threshold, and determines as the recommended financial product a financial product that satisfies either a first condition that the purchase score is greater than the first threshold and a second condition that the purchase amount is greater than the second threshold . The recommendation device of claim 1, wherein the determination unit determines the priority order of the recommended items in the following order: financial products that satisfy both the first condition and the second condition, financial products that satisfy the first condition but do not satisfy the second condition, and financial products that do not satisfy the first condition but do satisfy the second condition.

Images