JP7601848B2 - Information processing device, information processing method, information processing program, and learning model - Google Patents

Description

The present invention relates to an information processing device, an information processing method, an information processing program, and a learning model, and in particular to technology related to package delivery.

With the recent spread of electronic commerce, the demand for package delivery is rapidly increasing. If the user is not at home, the delivery company must redeliver the package, and even if the package is redelivered, there is no guarantee that the user will be at home, resulting in the need for further redelivery, which places a heavy burden on the delivery company. As a technology to address this issue, Patent Document 1 discloses a technology that predicts (determines) the possibility of a user being at home using a learning model that is trained using the relationship between the user's usage of the service and result information indicating whether the user was at home.

JP 2019-016411 A

The technology disclosed in Patent Document 1 predicts the likelihood of a target user being at home, i.e., the likelihood of successful delivery, based on the user's usage of the service. In this way, traditionally, the likelihood of successful delivery has been predicted based on user characteristics, such as the user's usage of the service. Meanwhile, the information required for delivery includes not only information about the destination user, but also information about the package's destination address, but until now, no mechanism has been provided for predicting the likelihood of successful delivery while taking into account characteristics related to the package's destination address.

The present invention was made in consideration of the above-mentioned problems, and aims to provide a mechanism for predicting the possibility of successful delivery to a destination user by taking into account information about the destination address of the package.

In order to solve the above problem, one aspect of an information processing device according to the present invention has a user feature acquisition means for acquiring user features representing the characteristics of a destination user of a package, a date feature acquisition means for acquiring date features representing the characteristics of the delivery date of the package, an identification means for identifying a building type of a building located at the destination address of the package, and a prediction means for predicting the possibility of successful delivery of the package to the destination user for each of a number of time periods on the delivery date based on the user features, the date features, and the building type.
In order to solve the above problem, another aspect of the information processing device according to the present invention includes a user feature acquisition means for acquiring user features representing characteristics of a destination user for each of a plurality of packages, a date feature acquisition means for acquiring date features representing characteristics of the delivery dates of the plurality of packages, an identification means for identifying a building type of a building located at a destination address for each of the plurality of packages, and a prediction means for predicting the possibility of successful delivery of each of the plurality of packages to the destination user for each of a plurality of time periods on the delivery date based on the user features, the date features, and the building type.

In order to solve the above problem, one aspect of the information processing method according to the present invention includes a user feature acquisition step of acquiring user features representative of the characteristics of the destination user of the package, a date feature acquisition step of acquiring date features representative of the characteristics of the delivery date of the package, an identification step of identifying the building type of the building located at the destination address of the package, and a prediction step of predicting the possibility of successful delivery of the package to the destination user for each of a number of time slots on the delivery date based on the user features, the date features, and the building type.

In order to solve the above problem, one aspect of the information processing program according to the present invention is an information processing program for causing a computer to execute information processing, the program causing the computer to execute processes including a user feature acquisition process for acquiring user features representative of the features of the destination user of the package, a date feature acquisition process for acquiring date features representative of the features of the delivery date of the package, an identification process for identifying the building type of the building located at the destination address of the package, and a prediction process for predicting the possibility of successful delivery of the package to the destination user for each of a number of time slots on the delivery date based on the user features, the date features, and the building type.

According to the present invention, a mechanism is provided for predicting the likelihood of successful delivery to a destination user by taking into account information regarding the destination address of a package.
The above-mentioned objects, aspects, and advantages of the present invention, as well as objects, aspects, and advantages of the present invention not described above, will be understood by those skilled in the art from the following detailed description of the invention by referring to the accompanying drawings and the claims.

FIG. 1 shows an example of the configuration of an information processing system. FIG. 2 shows an example of a functional configuration of the information processing device 10 according to the embodiment. FIG. 3 shows examples of addresses and building types identified by the addresses. FIG. 4 shows a conceptual diagram of the process of predicting the probability of successful delivery. FIG. 5 shows an example of a delivery table. FIG. 6A is a diagram for explaining the delivery schedule. FIG. 6B is a diagram for explaining the delivery schedule. FIG. 7 shows a flowchart of the entire process executed by the information processing device 10. FIG. 8 shows an example of the hardware configuration of the information processing device 10 and the user device 11.

Below, an embodiment for carrying out the present invention will be described in detail with reference to the attached drawings. Among the components disclosed below, those having the same functions are given the same reference numerals, and their description will be omitted. Note that the embodiment disclosed below is one example of a means for realizing the present invention, and should be appropriately modified or changed depending on the configuration of the device to which the present invention is applied and various conditions, and the present invention is not limited to the following embodiment. Furthermore, not all of the combinations of features described in this embodiment are necessarily essential to the solution of the present invention.

[Configuration of Information Processing System]
Fig. 1 shows an example of the configuration of an information processing system according to this embodiment. As an example, as shown in Fig. 1, this information processing system includes an information processing device 10 and multiple user devices 11-1 to 11-N (N>1) used by multiple arbitrary users 1 to N. In the following description, unless otherwise specified, the user devices 11-1 to 11-N may be collectively referred to as user device 11. In the following description, the terms user device and user may be used synonymously.

The user device 11 is, for example, a device such as a smartphone or a tablet, and is configured to be able to communicate with the information processing device 10 via a public network such as LTE (Long Term Evolution) or a wireless communication network such as a wireless LAN (Local Area Network). The user device 11 has a display unit (display surface) such as a liquid crystal display, and each user can perform various operations using a GUI (Graphic User Interface) equipped on the liquid crystal display. The operations include various operations on content such as images displayed on the screen, such as tapping, sliding, and scrolling using a finger or a stylus.
The user device 11 is not limited to the device shown in Fig. 1, but may be a desktop PC (Personal Computer) or a notebook PC. In this case, each user may operate the device using an input device such as a mouse or a keyboard. The user device 11 may also be provided with a separate display screen.

The user device 11 can log in to a web service (internet-related service) provided from the information processing device 10 or from another device (not shown) via the information processing device 10 to use the service. The web service can include an online mall, an online supermarket, or services related to communications, finance, real estate, sports, and travel, which are provided via the internet. By using such a web service, the user device 11 can transmit information about the user of the user device 11 to the information processing device 10.
For example, the user device 11 can transmit to the information processing device 10 information on characteristics of the user device and the user, such as the IP (Internet Protocol) address of the user device 11, the user's address, and the user's name.

In this embodiment, in order to use the web service, each of users 1 to N is assigned a user ID for identifying each user. The user ID may be an ID for identifying the user on the web service, or an ID for physically identifying the user device.

The information processing device 10 acquires various information from the user device 11, and also acquires delivery requests for packages addressed to multiple users, and determines a delivery schedule for the multiple packages on a specified date. In this embodiment, the information processing device 10 is configured to determine the delivery schedule taking into account the characteristics of the destination addresses of the multiple packages included in the delivery requests.

[Functional configuration of information processing device 10]
The information processing device 10 according to this embodiment acquires various information from the user devices 11-1 to 11-N. The information processing device 10 also acquires a delivery request for parcels addressed to multiple users. The delivery request includes a delivery date (a specified date) and information on multiple destination users and multiple destination addresses for the multiple parcels. Based on the acquired information, the information processing device 10 predicts the possibility of successful delivery in multiple time slots set for the delivery date for each of the multiple destination users. Then, based on the prediction result, the information processing device 10 determines an efficient delivery schedule for delivering parcels to multiple destination users on the delivery date. Note that the parcels may include any item as long as they can be delivered (delivered).

FIG. 2 shows an example of the functional configuration of the information processing device 10 according to the present embodiment.
The information processing device 10 shown in FIG. 2 includes a delivery request acquisition unit 101, a user feature acquisition unit 102, an order feature acquisition unit 103, a date feature acquisition unit 104, a building type identification unit 105, a delivery success prediction unit 106, a delivery schedule determination unit 107, an output unit 108, a learning model storage unit 110, a feature storage unit 120, and a rule storage unit 130. The learning model storage unit 110 stores a first time slot prediction model 111, a second time slot prediction model 112, a third time slot prediction model 113, a fourth time slot prediction model 114, a fifth time slot prediction model 115, a sixth time slot prediction model 116, and a user feature prediction model 117. The various learning models will be described later. In addition, the feature storage unit 120 stores a user feature 121, an order feature 122, a date feature 123, and a building type (building feature) 124, and the rule storage unit 130 stores a building type identification rule 131. The various features and rules will be described later.

The delivery request acquisition unit 101 acquires delivery requests for packages addressed to multiple users (two or more users among users 1 to N). The delivery request may be input by an operator via the input unit (input unit 85 in FIG. 8) or via the communication unit (communication I/F 87 in FIG. 8). Alternatively, the delivery request may be set in advance in the system, or may be set by any program stored in the memory unit (ROM 82 or RAM 83 in FIG. 8).

The delivery request includes information necessary for delivery for each of the multiple packages. In this embodiment, the delivery request includes information on multiple destination users, information on the addresses of the multiple destination users (multiple different destination addresses), and information on the delivery date, which is information on the date of scheduled delivery. In this embodiment, the delivery date is assumed to be the same date. In addition, the multiple destination users included in the delivery request are configured to be users whose addresses are close to each other (for example, the same city or town, or the delivery area of a single courier). In other words, the destination address information (multiple destination addresses) included in the delivery request may include information on multiple different addresses that satisfy one or more specified positional relationships. The delivery request may be configured in advance to include destination users whose addresses are close to each other, or the information processing device 10 or another device may generate a delivery request to include destination users whose addresses are close to each other.

The destination address information may include information on addresses located nearby (i.e., within a predetermined distance range). The destination address information may also include information on multiple different addresses located near a common road, street, or street. The destination address information may also include information on multiple addresses that have some in common, such as a common building name or house number. The destination address information may also include information on multiple different addresses located in a common block. The destination address information may also include multiple different addresses whose difference in altitude above sea level (height difference) is below a predetermined value. The destination address information may also include information on multiple different addresses whose number of right or left turns on the route between the addresses is below a predetermined number. The destination address information may also include information on multiple different addresses whose number of crossroads on the route between the addresses is below a predetermined number. The destination address information may also include information on multiple different addresses whose number of traffic lights on the route between the addresses is below a predetermined number (prediction values such as expected values of stop times due to traffic lights are below a predetermined time). The destination address information may also include information on multiple addresses where the road width of the road, street, or street that constitutes the route between the addresses exceeds a predetermined value. The destination address information may also include information on multiple addresses where the route between the addresses does not include a predetermined road, road, or street, such as a school zone.

The user feature acquisition unit 102 acquires user features of each of the multiple destination users included in the delivery request, among users 1 to N. In this embodiment, the user features are factual features (factual information) about the user device or user, and are fact-based features (information) that are actually or objectively obtained from the user device or user. The user feature acquisition unit 102 can acquire user features directly from the user device 11, for example. The user feature acquisition unit 102 can also acquire user features as information registered in a specified web service by the user of the user device 11. The user features are linked to the user ID of the destination user (hereinafter, destination user ID).

In this embodiment, the user characteristic acquisition unit 102 acquires demographic information and past purchasing behavior information in web services as user characteristics. Demographic information is information indicating demographic user attributes such as gender, age, residential area, occupation, and family structure. Purchasing behavior information is, for example, information on the name of the purchased item and the date and time of purchase. In addition, the user characteristic acquisition unit 102 may acquire data indicating the preferences registered by the user in the web service at the time of purchase as user characteristics.

The user feature acquisition unit 102 is also configured to apply the acquired user features to the trained user feature prediction model 117 to acquire user features (estimated user features (attributes)) estimated for the user features. For example, the user feature prediction model 117 is configured to receive user features (i.e., factual features) of a destination user (each user among a plurality of destination users) as input, and output the probability (correspondence probability) that each of the plurality of user features corresponds (matches) to the destination user. The user feature acquisition unit 102 ultimately determines the estimated user features of the destination user from the correlation probability.

For example, the user feature acquisition unit 102 inputs the demographic information of the destination user and data indicating past purchase history and preferences in web services of the destination user as the user features of the destination user to the user feature prediction model 117. The user feature prediction model 117 outputs, as the corresponding probability, the probability for each of a plurality of items that the destination user is estimated to purchase and a plurality of preferences that the destination user may have. Then, the user feature acquisition unit 102 can acquire items and preferences that have a probability equal to or greater than a predetermined value as the estimated user features of the destination user. In addition, the probability itself may be the estimated user feature.
In this embodiment, the estimated user characteristics include, for example, characteristics related to life stage, such as the probability of being working or having children, and characteristics related to lifestyle, such as the probability of living in a house or apartment (residence).

The user feature acquisition unit 102 stores the acquired and estimated user features as user features 121 in the feature storage unit 120. The user features are stored in association with the destination user ID.

The order feature acquisition unit 103 acquires the order features (order features) of each of the multiple destination users included in the delivery request, among users 1 to N. In this embodiment, the order features are information that represents the features of items (i.e., packages) previously ordered by the destination user. The ordered items may be limited to items ordered by the destination user and delivered to the destination user. The order features are obtained, for example, from the destination user's past order information (purchase behavior information) in a web service. For example, the order features include the weight of the item, the type of the item, and the type of genre to which the item belongs. The order features are linked to the destination user ID, and the order feature acquisition unit 103 stores the acquired order features in the feature storage unit 120 as the order features 122, linked to the destination user ID.

The date feature acquisition unit 104 acquires (identifies) the features of the date (date features) from the information on the delivery date of the package included in the delivery request. The date features can be acquired from calendar information, etc. The date feature acquisition unit 104 acquires, for example, the year, month, and day of the date and the type of the date (holiday, public holiday, or weekday) as date features, and stores them in the feature storage unit 120 as date features 123. The date features are stored in association with the destination user ID.

The building type identification unit 105 identifies the building type (building characteristics) of the building located at each address from the information of multiple different destination addresses included in the delivery request. In this embodiment, the building type identification unit 105 analyzes the character string of the destination address and identifies the building type of the address based on the building type identification rule 131 stored in the rule storage unit 130. In other words, the building type identification unit 105 applies the character string of the destination address to the building type identification rule 131 to identify the building type of the address. The procedure for identifying the building type will be described later.

Furthermore, the building type identification unit 105 can determine, based on the destination address information, whether or not the building located at the address has a delivery box (a box for storing delivered packages). This determination can be made using an online public or private service, such as a specified web service. Based on this determination, the building type identification unit 105 generates delivery box information indicating whether or not the building located at the destination address has a delivery box.

Note that a delivery box in a building located at a destination address can be understood to be synonymous with a stationary or mobile storage box in the building that can accommodate delivered packages. The presence of a stationary or mobile storage box can also be determined using online public or private services.

The building type identification unit 105 links the identified building type to the destination user ID and stores it in the feature storage unit 120 as the building type 124. Here, the building type identification unit 105 may link the generated delivery box information (information indicating whether the building has a delivery box or not) to the identified building type and store it as the building type 124.

The delivery success prediction unit 106 predicts the possibility of successful delivery to each of the multiple destination users included in the delivery request based on the user features 121, order features 122, date features 123, and building type 124. The delivery success prediction unit 106 may predict the possibility of successful delivery based on the user features 121, date features 123, and building type 124. In this embodiment, the delivery success prediction unit 106 predicts the possibility of successful delivery for each of multiple time slots (time slots) on the delivery date. The procedure for predicting the possibility of successful delivery will be described later.

The delivery schedule determination unit 107 determines a delivery schedule for the delivery date based on the possibility of successful delivery to each of the multiple destination users predicted by the delivery success prediction unit 106. In this embodiment, the delivery schedule includes the time slot (time period) for delivery, and information on the destination user and destination address. The procedure for determining the delivery schedule will be described later.

The output unit 108 outputs the delivery schedule determined by the delivery schedule determination unit 107. The output unit 108 may generate and output information related to the schedule. The output may be any output process, and may be output to an external device via a communication I/F (communication I/F 87 in FIG. 8) or may be displayed on a display unit (display unit 86 in FIG. 8).

[Procedure for identifying building types]
The following describes the procedure of the building type identification process performed by the building type identification unit 105. In this embodiment, the building type identification unit 105 identifies a building type based on the building type identification rules 131 stored in the rule storage unit 130. Specifically, the building type identification unit 105 applies the character string of the address to the building type identification rules 131 to identify the building type.

The building type identification rule 131 is a pre-created rule for identifying a building type based on the characteristics of an address character string. The building type identification rule 131 may be input by an operator via an input unit (input unit 85 in FIG. 8 ) or via a communication unit (communication I/F 87 in FIG. 8 ). Alternatively, the building type identification rule 131 may be set in advance in the system, or may be set by an arbitrary program stored in a storage unit (ROM 82 or RAM 83 in FIG. 8 ).
Note that the building type identification rules 131 described in this embodiment are merely an example, and although there may be cases where they do not match actual examples of combinations of addresses and building types, any rule capable of identifying the building type of an address from the character string of the address can be used. For example, the building type may be identified depending on whether or not the character string of the address contains a predetermined keyword. Furthermore, new rules may be added to the building type identification rules 131.

The building type identification rule 131 according to this embodiment has nine types of rules, from (1) to (9). The rules (1) to (9) are configured to identify a building type based on the characteristics of a character string of a destination address.
(1) If the address includes "Inc.", the building type of the address is identified as "business building," which refers to a business or company building.
For example, "1-2-3 D-cho, ABC-shi, Osaka Prefecture, inside abc Co., Ltd." is identified as a "business building."

(2) If the address includes the word "building" followed by three digits, the building type of the address is identified as a "unit building," which refers to a building, condominium, or apartment (residence) with multiple residential units (individually occupied spaces).
For example, "abc building 303, 1-2-3, D-cho, ABC-ku, Tokyo" is identified as a "unit building."

(3) If an address includes the word "building" but does not include three digits following the word "building," the building type of the address is identified as "business building."
For example, "abc building, 5th floor, 1-2-3 D-cho, ABC-ku, Tokyo" is identified as a "business building."

(4) If the address includes the final character "number", the building type of the address is identified as "unit building".
For example, "Residence abc 303, 1-2-3, D-cho, ABC-shi, Osaka Prefecture" is identified as a "unit building."

(5) If an address includes an “x1-x2-x3” format, where “x3” is not a three or four digit number, the building type of the address is identified as “home (private residence, detached house).”
Example: "1-234-5 D-cho, ABC City, Saitama Prefecture" is identified as a "house."

(6) If the address contains an “x1-x2” format, the building type of the address is identified as “unit building.”
For example, "DE123-4, ABC-cho, Kanagawa Prefecture" is identified as a "unit building."

(7) If the address contains a single digit and it is at the end, the building type of the address is identified as "home."
For example, "1234 D-cho, ABC-shi, Chiba Prefecture" is identified as a "house."

(8) If an address contains an address in the format “x1-x2-x3” or “x1-x2-x3-x4”, where “x3” or “x4” is a three or four digit number, the building type of the address is identified as “unit building”.
For example, "FG1-2-3-303, DE ward, ABC city, Fukuoka prefecture" is identified as a "unit building."

(9) If an address contains an address in the format “x1-x2-x3” or “x1-x2-x3-x4” and contains a number at the end in addition to the format, the building type of the address is identified as a “unit building.”
For example, "Residence abc 303, DE1-2-3, ABC City, Kanagawa Prefecture" is identified as a "unit building."

In addition, the building type identification unit 105 can be configured to recognize "x1-x2-x3" in the above rules as the same string as variations such as "x1-chome x2-no. x3" or "x1-chome x2-x3."
In addition, the building type identification unit 105 can be configured to recognize, in the above rules, "x1-x2-x3-x4" as the same string as variations such as "x1-chome x no. 2 x no. 3 (building name) x 4" or "x1-chome x 2-x3 (building name) x 4."

Figure 3 shows a list of examples of addresses listed in rules (1) to (9) above and the building types identified by the addresses. As shown in Figure 3, the destination address is classified into one of the building types "business building," "unit building," and "house" according to the characteristics of the character string contained in the address.

For an address whose building type cannot be identified by the above rules (1) to (9), the building type identification unit 105 may identify the building type of the address as a predetermined building type such as "house." Alternatively, for an address whose building type cannot be identified by the above rules (1) to (9), the building type identification unit 105 may identify the building type of the address as "unknown type."

In addition, rules (1) to (9) may be applicable to some areas, such as Tokyo, but may not be applicable to other areas. Therefore, rules other than rules (1) to (9) may be created and applied to other areas. Furthermore, rules other than rules (1) to (9) may be created and applied even to areas within Tokyo.

[Procedure for predicting the probability of successful delivery]
The procedure for predicting the possibility of successful delivery to a destination user by the delivery success prediction unit 106 will be described. In this embodiment, the delivery success prediction unit 106 predicts the possibility of successful delivery to a destination user using a learning model for trained machine learning. In this embodiment, the delivery time is divided into six time slots (time slots), namely, 8:00 to 12:00 (first time slot), 12:00 to 14:00 (second time slot), 14:00 to 16:00 (third time slot), 16:00 to 18:00 (fourth time slot), 18:00 to 20:00 (fifth time slot), and 20:00 to 22:00 (sixth time slot), and the possibility of successful delivery for each slot is predicted.

The first time slot prediction model 111 through the sixth time slot prediction model 116 are used to predict the probability of successful delivery.
The first time slot prediction model 111 is a learning model that predicts the probability of successful delivery from 8:00 to 12:00 (first time slot).
The second time slot prediction model 112 is a learning model that predicts the probability of successful delivery from 12:00 to 14:00 (the second time slot).
The third time slot prediction model 113 is a learning model that predicts the probability of successful delivery from 14:00 to 16:00 (the third time slot).
The fourth time slot prediction model 114 is a learning model that predicts the probability of successful delivery from 4:00 p.m. to 6:00 p.m. (fourth time slot).
The fifth time slot prediction model 115 is a learning model that predicts the probability of successful delivery from 18:00 to 20:00 (the fifth time slot).
The sixth time slot prediction model 116 is a learning model that predicts the probability of successful delivery from 8 p.m. to 10 p.m. (the sixth time slot).

The first time slot prediction model 111 to the sixth time slot prediction model 116 are learning models for machine learning, for example, learning models based on CatBoost. Alternatively, the first time slot prediction model 111 to the sixth time slot prediction model 116 may be other boosting-based learning models such as XGBoost or LightGBM.

In the learning stage, a learning model corresponding to each time slot is trained using a combination of user features 121, order features 122, date features 123, and delivery history (correct answer data (label)) linked to at least one of building type 124 stored in feature storage unit 120 for past deliveries to users 1 to N as training data. The delivery history indicates successful delivery (=1) or failed delivery (=0) for each time slot. Successful delivery may include, for example, the presence of the destination user in the building of the destination address, or the delivery to a delivery box provided in the building. Alternatively, successful delivery may include completion of delivery without meeting the user (for example, leaving the package at the entrance of each unit of the building or at the front door of the house). Failed delivery may include, for example, the absence of a user in the building of the destination address, or the failure to deliver to a delivery box provided in the building.

For example, in a past delivery, if delivery to user 1 was completed at 11:00, the delivery history of only the first time slot among the first to sixth time slots will indicate delivery success (=1), and the delivery history of the other time slots will indicate delivery failure (=0). In addition, as input data, at least one of user characteristics 121 of user 1, order characteristics 122 related to the items ordered by user 1 for the delivery, date characteristics 123 of the delivery date, and building type 124 of the delivery destination address (which may include delivery box information) is used.
The first time slot prediction model 111 is trained using training data that is a combination of the input data and delivery success (=1), while the second time slot prediction model to the sixth time slot prediction model 116 are trained using training data that is a combination of the input data and delivery failure (=0).

The trained first time slot prediction model 111 is configured to receive as input the user features 121 and order features 122 of any target user, the date features 123 of the delivery date, and the building type 124 of the delivery address, and to output (predict) the possibility of successful delivery to the target user in the first time slot. Alternatively, the first time slot prediction model 111 may be configured to receive as input the user features 121, the date features 123, and the building type 124, and to output the possibility of successful delivery to the target user in the first time slot.
Similarly, the second time slot prediction model 112 to the sixth time slot prediction model 116 are each configured to predict the probability of successful delivery to a target user in the second time slot to the sixth time slot, respectively, using the user features 121 and order features 122 of a target user, the date features 123 of a delivery date, and the building type 124 of a delivery address as input. Alternatively, the second time slot prediction model 112 to the sixth time slot prediction model 116 may each be configured to output the probability of successful delivery to the target user in the first time slot to the sixth time slot, respectively, using the user features 121, the date features 123, and the building type 124 as input.
The first time slot prediction model 111 through the sixth time slot prediction model 116 are configured to output the probability of successful transmission in each time slot as a number between 0 and 1 (1 being the maximum probability).

In the prediction stage, the probability of successful delivery to each of the multiple destination users included in the delivery request acquired by the delivery request acquisition unit 101 is predicted for each time slot.
4 shows a conceptual diagram of the process of predicting the probability of successful delivery. The delivery success prediction unit 106 predicts the probability of successful delivery to each of the multiple destination users included in the delivery request acquired by the delivery request acquisition unit 101 for the first to sixth time slots.

Specifically, the delivery success prediction unit 106 prepares input data 40 including, for each of the multiple destination users, a user feature 121, an order feature 122, a date feature 123 of the delivery date, and a building type 124 of the delivery address. Note that the order feature 122 does not have to be included in the input data 40.
The user characteristics 121 are the user characteristics of the destination user included in the delivery request, and are acquired by the user characteristic acquisition unit 102 .
The order characteristics 122 are the order characteristics of the destination user included in the delivery request, and are acquired by the order characteristics acquisition unit 103 .
The date feature 123 is a date feature of the delivery date included in the delivery request, and is acquired (specified) by the date feature specification unit 194. In this embodiment, it is assumed that the delivery dates included in the delivery request (i.e., the delivery dates for multiple packages) are the same date.
The building type 124 is the building type of the building located at the destination address included in the delivery request, and is identified by the building type identification unit 105. The building type 124 may also include delivery box information (information indicating whether the building located at the destination address has a delivery box or not).

Then, the delivery success prediction unit 106 inputs the input data 40 from the first time slot prediction model 111 to the sixth time slot prediction model 116, and predicts the possibility of delivery success in the first time slot to the sixth time slot for each of the multiple destination users included in the delivery request.

The first time slot prediction model 111 outputs a delivery success probability 41 (a number between 0 and 1) for the destination user in the first time slot.
The second time slot prediction model 112 outputs the destination user's probability of successful delivery 42 in the second time slot (a number between 0 and 1).
The third time slot prediction model 113 outputs the destination user's probability of successful delivery in the third time slot 43 (a number between 0 and 1).
The fourth time slot prediction model 114 outputs the destination user's probability of successful delivery in the fourth time slot 44 (a number between 0 and 1).
The 5th time slot prediction model 115 outputs the destination user's probability of successful delivery in the 5th time slot 45 (a number between 0 and 1).
The sixth time slot prediction model 116 outputs the destination user's probability of successful delivery in the sixth time slot 46 (a number between 0 and 1).

The delivery success prediction unit 106 determines whether delivery is possible or not possible for each destination user for each time slot based on the delivery success probabilities 41 to 46 output for all multiple destination users included in the delivery request. For example, the delivery success prediction unit 106 determines "OK (delivery possible)" for time slots with a delivery success probability equal to or greater than a predetermined threshold. On the other hand, the delivery success prediction unit 106 determines "NG (delivery not possible)" for time slots with a delivery success probability less than the predetermined threshold. The predetermined threshold is, for example, 0.5. The delivery success prediction unit 106 creates a delivery table including the "OK" or "NG" determined for each time slot for each destination user.

The delivery table 47 shown in FIG. 4 is an example of a delivery table created by the delivery success prediction unit 106. In the delivery table 47, a label (indicator) of "OK (delivery possible)" or "NG (delivery not possible)" is assigned to each time slot for any of the multiple destination users included in the delivery request. The delivery success prediction unit 106 creates such a delivery table for all of the multiple destination users.

Figure 5 shows an example of a delivery table when the multiple destination users included in a delivery request are users 1 to 4. Delivery tables T1 to T4 are created based on the predicted probability of successful delivery for each time slot for each of users 1 to 4, and a label (indicator) of "OK (delivery possible)" or "NG (delivery not possible)" is attached to each time slot.

In this embodiment, one threshold is set, and a delivery table is created that includes delivery possible or impossible determined using the threshold from the predicted delivery success probability, but two or more thresholds may be set. For example, when two thresholds, a first threshold (=0.8) and a second threshold (=0.4), are set, a delivery table including three labels (indicators) of "probability of success: high", "probability of success: medium", and "probability of success: low" may be created. That is, for a time slot with a delivery success probability equal to or greater than the first threshold, it may be determined that "probability of success: high", for a time slot with a delivery success probability equal to or greater than the second threshold and less than the first threshold, it may be determined that "probability of success: medium", and for a time slot with a delivery success probability less than the second threshold, it may be determined that "probability of success: low".
Alternatively, the delivery table may use a predicted probability of successful delivery (a number between 0 and 1), or a percentage (%) of the predicted probability of successful delivery, in which case the number directly represents the probability of successful delivery.

The trained model such as the first time slot prediction model 111 may be trained using teacher data in which various data except the building type 124 is labeled, and may output the delivery success probability (a value from 0 to 1) by inputting the input data 40 except the building type 124. At this time, the delivery success prediction unit 106 may apply some coefficient to the delivery success probability output by the trained model according to the building type 124. For example, if the building type 124 is a business building, the delivery success prediction unit 106 may apply some coefficient to upwardly revise the delivery success probability output by the trained model. In addition, the trained model may be trained using teacher data in which various data except the date feature 123 is labeled, and may output the delivery success probability by inputting the input data 40 except the date feature 123. At this time, the delivery success prediction unit 106 may apply some coefficient to the delivery success probability output by the trained model according to the date feature 123 and the type of time slot. For example, if the date feature 123 indicates a weekday that may be a business day, and the time slot type indicates daytime that may be within business hours, the delivery success prediction unit 106 may apply some coefficient to upwardly revise the delivery success probability output by the trained model.

[Delivery schedule determination procedure]
When the delivery success prediction unit 106 predicts the delivery success probability for each time slot for all of the multiple destination users and creates a delivery table, the delivery schedule determination unit 107 determines a delivery schedule. As described above, in this embodiment, the delivery schedule includes information on the time slot (time period) for delivery, the destination user, and the destination address.

The delivery schedule determination unit 107 extracts (groups) two or more users who share the nth (in this embodiment, 1≦n≦6) time slot indicating OK in the delivery table created by the delivery success prediction unit 106. Then, the delivery schedule determination unit 107 determines a delivery schedule so that packages addressed to the two or more extracted (grouped) users are delivered in the nth time slot. Note that if it is not possible to extract two or more users who share the nth time slot, the delivery schedule is determined so that delivery can be made in the time slot indicating OK.

An example of a delivery schedule that is determined when delivery tables T1 to T4 for users 1 to 4 are created as shown in FIG. 5 will be described with reference to FIG. 6A and FIG. 6B. FIG. 6A and FIG. 6B are diagrams for explaining a delivery schedule. In FIG. 6A and FIG. 6B, delivery tables T1 to T4 indicate delivery tables for users 1 to 4 as shown in FIG. 5, and destinations D1 to D4 indicate the destination addresses of users 1 to 4.

The delivery schedule determination unit 107 refers to delivery tables T1 to T4, recognizes that the first time slot is both OK in delivery tables T1 and T2, and decides to deliver packages addressed to users 1 and 2 in the first time slot. Note that, since the second time slot is also both OK in delivery tables T1 and T2, the delivery schedule determination unit 107 may decide to deliver packages addressed to users 1 and 2 between the first and second time slots.
In addition, the delivery schedule determination unit 107 refers to delivery tables T1 to T4, recognizes that the fourth to sixth time slots both indicate OK in delivery tables T3 and T4, and determines that packages addressed to users 3 and 4 will be delivered between the fourth and sixth time slots.

Therefore, the delivery schedule determination unit 107 determines the following delivery schedule, as an example.
1st time slot: User 1 (destination D1)
User 2 (Destination D2)
4th to 6th time slots: User 3 (destination D3)
User 4 (Destination D4)

6A illustrates a delivery status in the first time slot according to the determined delivery schedule, specifically showing that packages addressed to user 1 and user 2 are delivered in the first time slot. The delivery order for user 1 and user 2 can be determined arbitrarily. Alternatively, the delivery order can be determined based on the location of the delivery vehicle (or delivery person) and the locations of destinations D1 and D2 so that delivery can be made via an efficient route.
6B illustrates the delivery status in the fourth to sixth time slots according to the determined delivery schedule, specifically showing that packages addressed to user 3 and user 4 are delivered in the fourth to sixth time slots. The delivery order for user 3 and user 4 can be determined arbitrarily. Alternatively, the delivery order can be determined based on the location of the delivery vehicle (or delivery person) and the locations of destinations D3 and D4 so as to enable delivery via an efficient route.

In this way, multiple destination users who have a high probability of successful delivery in the same time slot are grouped together, and a delivery schedule is determined so that delivery is made to the multiple destination users in that group in that time slot. Then, delivery is made according to that delivery schedule, which makes it possible to reduce the possibility of delivery failure and to make efficient delivery.

6A and 6B, if delivery is attempted from user 1 to user 4 in the first time slot, delivery to users 3 and 4 is likely to fail, which may result in redelivery to users 3 and 4. Furthermore, if redelivery is attempted to users 3 and 4 in the third time slot, delivery is likely to fail again, which may increase the number of redeliveries.
On the other hand, according to this embodiment, the delivery schedule is determined so that deliveries are made to multiple users who have a high probability of successful delivery in the same time slot, thereby making it possible to reduce the number of redeliveries.

[Overall processing flow]
7 shows a flowchart of the overall process executed by the information processing device 10 according to this embodiment. For the explanation of FIG. 7, the information processing system shown in FIG.

In S71, the delivery request acquisition unit 101 acquires a delivery request for a package addressed to multiple users (destination users) among users 1 to N. The delivery request includes information on the multiple destination users, information on the addresses of each of the multiple destination users (multiple different destination addresses), and information on the delivery date.

In S72, the user feature acquisition unit 102 acquires the user features of each of the multiple destination users and stores them in the feature storage unit 120 as user features 121. The process of S72 may be a process of acquiring (collecting) user features and estimated user features for a certain period of time in the past. The user features 121 are linked to the destination user ID and stored in the feature storage unit 120.

In S<b>73 , the order feature acquisition unit 103 acquires the order features of each of the multiple destination users, and stores them in the feature storage unit 120 as order features 122 .
In S74, the date feature acquisition unit 104 acquires the date feature of the delivery date, and stores it in the feature storage unit 120 as a date feature 123. The order feature 122 and the date feature 123 are stored in the feature storage unit 120 in association with the destination user ID.

In S75, the building type identification unit 105 identifies the building type of the building located at the destination address from the destination address information. The procedure for identifying the building type is as described above. As an option, the building type identification unit 105 determines whether the building located at the destination address is equipped with a delivery box and generates delivery box information. The building type identification unit 105 stores the identified building type in the feature storage unit 120 as the building type 124. Furthermore, when delivery box information is generated, the delivery box information is included in the building type 124. The building type 124 is associated with the destination user ID and stored in the feature storage unit 120.
The order of the processes from S72 to S75 is not limited to the order shown in FIG. 7, and the processes may be performed in a different order or simultaneously.

In S76, the delivery success prediction unit 106 predicts the possibility of successful delivery in multiple time slots for multiple destination users. The prediction is performed by applying input data including the user features 121, order features 122, date features 123, and building type 124 acquired or identified in S72 to S75 to the first time slot prediction model 111 to the sixth time slot prediction model 116. The input data does not have to include the order features 122. The procedure for predicting the possibility of successful delivery is as described above with reference to FIG. 4. After predicting the possibility of successful delivery, the delivery success prediction unit 106 creates a delivery table for each of the multiple destination users. Examples of the delivery table are shown in FIG. 4 and FIG. 5.

In S77, the delivery schedule determination unit 107 determines a delivery schedule based on the delivery table created in S76. In the case of the delivery tables shown in FIG. 4 and FIG. 5, the delivery schedule determination unit 107 extracts two or more users who share the n-th (in this embodiment, 1≦n≦6) time slot indicating OK. The delivery schedule determination unit 107 then determines a delivery schedule so that packages addressed to the two or more extracted users are delivered in the n-th time slot. In this embodiment, since the date information included in the delivery request acquired in S71 is the same date, a delivery schedule is determined in which the destination user and destination address are set for each time slot on that date.

In S78, the output unit 108 outputs the delivery schedule determined (created) in S77. The output unit 108 may generate information related to the delivery schedule and output it to an external device (not shown). The information related to the delivery schedule is, for example, information such as the position information of each household on a map and the predicted arrival time at each household based on the road congestion situation. The position information can be identified from address information.

[Hardware configuration of information processing device 10]
FIG. 8 is a block diagram showing an example of a hardware configuration of the information processing device 10 according to the present embodiment.
The information processing apparatus 10 according to the present embodiment can be implemented on any single or multiple computers, mobile devices, or any other processing platform.
8, the information processing device 10 is illustrated as being implemented in a single computer, but the information processing device 10 according to the present embodiment may be implemented in a computer system including multiple computers. The multiple computers may be connected to each other via a wired or wireless network so as to be able to communicate with each other.

8, the information processing device 10 may include a CPU 81, a ROM 82, a RAM 83, a HDD 84, an input unit 85, a display unit 86, a communication I/F 87, and a system bus 88. The information processing device 10 may also include an external memory.
A CPU (Central Processing Unit) 81 generally controls the operations of the information processing device 10, and controls each of the components (82 to 87) via a system bus 88, which is a data transmission path.

The ROM (Read Only Memory) 82 is a non-volatile memory that stores control programs and the like necessary for the CPU 81 to execute processing. Note that the programs may be stored in a non-volatile memory such as a HDD (Hard Disk Drive) 84 or an SSD (Solid State Drive) or in an external memory such as a removable storage medium (not shown).
The RAM (Random Access Memory) 83 is a volatile memory and functions as a main memory, a work area, etc. of the CPU 81. That is, when executing a process, the CPU 81 loads necessary programs, etc. from the ROM 82 into the RAM 83 and executes the programs, etc. to realize various functional operations. The learning model storage unit 110, the feature storage unit 120, and the rule storage unit 130 shown in FIG. 2 can be configured by the RAM 83.

The HDD 84 stores, for example, various data and various information required when the CPU 81 performs processing using a program. The HDD 84 also stores, for example, various data and various information obtained when the CPU 81 performs processing using a program.
The input unit 85 is composed of a keyboard and a pointing device such as a mouse.
The display unit 86 is configured with a monitor such as a liquid crystal display (LCD) etc. The display unit 86 may be configured in combination with the input unit 85 to function as a GUI (Graphical User Interface).

The communication I/F 87 is an interface that controls communication between the information processing device 10 and an external device.
The communication I/F 87 provides an interface with a network and executes communication with an external device via the network. Various data, various parameters, etc. are transmitted and received between the external device and the communication I/F 87. In this embodiment, the communication I/F 87 may execute communication via a wired LAN (Local Area Network) or a dedicated line that complies with a communication standard such as Ethernet (registered trademark). However, the network that can be used in this embodiment is not limited to this, and may be configured as a wireless network. This wireless network includes wireless PANs (Personal Area Networks) such as Bluetooth (registered trademark), ZigBee (registered trademark), and UWB (Ultra Wide Band). It also includes wireless LANs (Local Area Networks) such as Wi-Fi (Wireless Fidelity) (registered trademark) and wireless MANs (Metropolitan Area Networks) such as WiMAX (registered trademark). It also includes wireless WANs (Wide Area Networks) such as LTE/3G, 4G, and 5G. Note that the network only needs to be able to connect devices to each other and communicate with each other, and the communication standard, scale, and configuration are not limited to those described above.

At least some of the functions of each element of the information processing device 10 shown in FIG. 2 can be realized by the CPU 81 executing a program. However, at least some of the functions of each element of the information processing device 10 shown in FIG. 2 may be operated as dedicated hardware. In this case, the dedicated hardware operates under the control of the CPU 81.

[Hardware Configuration of User Device 11]
The hardware configuration of the user device 11 shown in Fig. 1 may be the same as that of Fig. 8. That is, the user device 11 may include a CPU 81, a ROM 82, a RAM 83, a HDD 84, an input unit 85, a display unit 86, a communication I/F 87, and a system bus 88. The user device 11 can display various information provided by the information processing device 10 on the display unit 86 and perform processing corresponding to an input operation received from a user via a GUI (configured by an input unit 95 and the display unit 86).

In this way, the information processing device 10 predicts the possibility of successful delivery to the destination user for each of a plurality of time slots (time periods) by using a learning model for machine learning, taking into account the user characteristics of the destination user and the characteristics of the destination address. By taking into account not only the user characteristics of the destination user but also the building type of the building located at the address, which is a characteristic of the destination address, the prediction accuracy of the possibility of successful delivery can be improved.
For example, in the learning stage, if the building type is a "house", the learning model may be trained using a non-face-to-face delivery history (successful delivery). Furthermore, the building type may include the presence or absence of a delivery box, and the learning model may be trained using a delivery history (successful delivery) to the delivery box.
Furthermore, in the prediction stage, even if the building type of the building at the destination address is a "business building," if the destination user lives in that building and the user characteristics include "lives in an apartment (residence)," it is possible to predict the possibility of successful delivery based on the user characteristics.
This allows the learning model to improve its prediction accuracy of the probability of successful delivery by taking into account the user characteristics of the destination user and the characteristics of the destination address.

In addition, the information processing device 10 groups multiple destination users who have a high probability of successful delivery in the same time slot (one or more time slots), and determines a delivery schedule so that delivery is made to the multiple grouped destination users in that time slot. Then, by making delivery according to the delivery schedule, the number of redeliveries can be reduced, making it possible to cut costs associated with redelivery.

In this embodiment, it is assumed that the delivery dates included in the delivery requests are the same date, but if different delivery dates are included, the delivery schedule determination unit 107 can determine a delivery schedule for each delivery date.

In addition, the first to sixth time slots set in this embodiment are merely examples, and the time width of each time slot may be determined arbitrarily, and the first to Mth time slots may be appropriately adopted (M>1). For example, if a building is equipped with delivery boxes, the time slots can be set at any time during the 24 hours.
In addition, in this embodiment, multiple learning models, such as six learning models, are used to predict the probability of successful delivery in the first to sixth time slots, but a single learning model that integrates the multiple learning models may also be used.

In addition, in this embodiment, it is assumed that the delivery request is configured to include destination users whose addresses are close to each other (for example, in the same city or town, or within the delivery area of a single courier delivery service), but the range of addresses of destination users included in the delivery request may be wider. In this case, for example, the information processing device 10 may predict the possibility of successful delivery to multiple destination users according to the above-mentioned procedure, and determine a delivery schedule for the delivery area of a single courier delivery service according to the prediction result.

Although specific embodiments have been described above, these embodiments are merely examples and are not intended to limit the scope of the present invention. The devices and methods described herein may be embodied in forms other than those described above. Furthermore, the above-described embodiments may be omitted, substituted, or modified as appropriate without departing from the scope of the present invention. Forms in which such omissions, substitutions, or modifications have been made are included within the scope of the claims and their equivalents, and belong to the technical scope of the present invention.

The disclosure of this embodiment includes the following configuration.
[1] An information processing device having a user feature acquisition means for acquiring user features representing the characteristics of a destination user of a package, a date feature acquisition means for acquiring date features representing the characteristics of the delivery date of the package, an identification means for identifying a building type of a building located at the destination address of the package, and a prediction means for predicting the possibility of successful delivery of the package to the destination user for each of a number of time periods on the delivery date based on the user features, the date features, and the building type.

[2] An information processing device having a user feature acquisition means for acquiring user features representing the features of a destination user for each of a plurality of packages, a date feature acquisition means for acquiring date features representing the features of the delivery dates of the plurality of packages, an identification means for identifying a building type of a building located at a destination address for each of the plurality of packages, and a prediction means for predicting the possibility of successful delivery of each of the plurality of packages to the destination user for each of a plurality of time periods on the delivery date based on the user features, the date features, and the building type.

[3] The information processing device described in [1] or [2], wherein the identification means identifies the building type based on a character string of the destination address.

[4] The information processing device described in [3], in which the identification means identifies the building type by applying a character string of the destination address to a predetermined rule.

[5] The information processing device according to any one of [1] to [4], wherein the identification means acquires delivery box information indicating whether or not the building located at the destination address has a delivery box, and includes the delivery box information in the building type.

[6] An information processing device according to any one of [1] to [5], further comprising an order feature acquisition means for acquiring order features representing the characteristics of items previously ordered by the destination user, and the prediction means for predicting the probability of successful delivery based on the user features, the date features, the building type, and the order features.

[7] The information processing device described in any one of [1] to [5], wherein the prediction means predicts the probability of successful delivery using a learning model for machine learning, and the learning model is configured to input the user features, the date features, and the building type, and output the probability of successful delivery for each of the multiple time periods.

[8] The information processing device described in [2] further includes a determination means for determining a delivery schedule for the multiple packages based on the predicted probability of successful delivery of each of the multiple packages to the destination user for each of the multiple time periods.

[9] A learning model that is trained to output the probability of successful delivery of the package to the destination user for each of multiple time periods on the delivery date when user features representing the characteristics of the destination user of the package, date features representing the characteristics of the delivery date of the package, and building features representing the building type of the building located at the destination address of the package are input.

[10] An information processing method comprising: a user feature acquisition step for acquiring user features representative of the characteristics of a destination user of a package; a date feature acquisition step for acquiring date features representative of the characteristics of the delivery date of the package; an identification step for identifying the building type of a building located at the destination address of the package; and a prediction step for predicting the possibility of successful delivery of the package to the destination user for each of a plurality of time slots on the delivery date based on the user features, the date features, and the building type.

[11] An information processing program for causing a computer to execute information processing, the program causing the computer to execute processes including a user feature acquisition process for acquiring user features representative of the features of a destination user of the parcel, a date feature acquisition process for acquiring date features representative of the features of the delivery date of the parcel, an identification process for identifying the building type of a building located at the destination address of the parcel, and a prediction process for predicting the possibility of successful delivery of the parcel to the destination user for each of a number of time slots on the delivery date based on the user features, the date features, and the building type.

1 to N: User, 10: Information processing device, 11-1 to 11-N: User device, 101: Delivery request acquisition unit, 102: User feature acquisition unit, 103: Order feature acquisition unit, 104: Date feature acquisition unit, 105: Building type identification unit, 106: Delivery success prediction unit, 107: Delivery schedule determination unit, 108: Output unit, 110: Learning model storage unit, 111: First time slot prediction model, 112: Second time slot prediction model, 113: Third time slot prediction model, 114: Fourth time slot prediction model, 115: Fifth time slot prediction model, 116: Sixth time slot prediction model, 117: User feature prediction model, 120: Feature storage unit, 121: User feature, 122: Order feature, 123: Date feature, 124: Building type, 130: Rule storage unit, 131: Building type identification rule

Claims (10)

A user characteristic acquisition means for acquiring a user characteristic representing a characteristic of a destination user of a package from a user device of the destination user ;
A date characteristic acquisition means for acquiring a date characteristic representing a characteristic of a delivery date of the package;
A means for identifying the building type of the building located at the destination address of the package as one of "business building" which refers to a business or company building, "unit building" which refers to a building, condominium, or apartment having multiple living spaces, "house," or "unknown" ;
a prediction means for predicting a probability of successful delivery of the package to the destination user for each of a plurality of time slots on the delivery date based on the user characteristics, the date characteristics, and the building type;
having
The user features include factual features of the destination user and estimated features estimated from the factual features using a first learning model for machine learning;
The information processing device is characterized in that the identification means identifies the building type based on predetermined rules regarding one or more predetermined characters contained in the string of the destination address, the position of the one or more characters in the string, the position and number of digits of the numbers contained in the string, and the notation format of the numbers contained in the string . a user characteristic acquisition means for acquiring user characteristics representing characteristics of a destination user of each of the plurality of packages from a user device of the destination user ;
A date feature acquiring means for acquiring date features representing the characteristics of the delivery dates of the plurality of packages;
A means for identifying the building type of the building located at the destination address of each of the plurality of packages as one of a "business building" which refers to a business or company building, a "unit building" which refers to a building, condominium, or apartment having multiple living spaces, a "house," or an "unknown type" ;
a prediction means for predicting a probability of successful delivery of each of the plurality of packages to a destination user for each of a plurality of time slots on the delivery date based on the user characteristics, the date characteristics, and the building type;
having
The user features include factual features of the destination user and estimated features of the destination user estimated from the factual features using a first learning model for machine learning;
The information processing device is characterized in that the identification means identifies the building type based on predetermined rules regarding the numbers and non-numeric parts contained in the string of the destination address, the position of the numbers contained in the string, the number of numeric digits contained in the string, the position of a predetermined kanji character in the string, and the notation format of the numbers contained in the string . The information processing device according to claim 1 or 2, characterized in that the factual characteristics of the destination user include demographic information of the destination user, a purchase history of the destination user in a web service, and data registered in the web service by the destination user, and the estimated characteristics of the destination user include characteristics estimated from the factual characteristics of the destination user using the first learning model. The specification means acquires delivery box information indicating whether or not a building located at the destination address has a delivery box, and includes the delivery box information in the building type.
3. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer. The method further includes an order characteristic acquisition means for acquiring order characteristics representing characteristics of items previously ordered by the destination user,
The prediction means predicts the likelihood of successful delivery based on the user characteristics, the date characteristics, the building type, and the order characteristics.
3. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer. The prediction means predicts the probability of successful delivery using a second learning model for machine learning;
The second learning model is configured to output the probability of successful delivery for each of the plurality of time periods using the user features, the date features, and the building type as inputs.
3. The information processing apparatus according to claim 1, wherein the information processing apparatus is a computer. and determining means for determining a delivery schedule for the plurality of packages based on the predicted probability of successful delivery of each of the plurality of packages to a destination user for each of the plurality of time periods.
3. The information processing apparatus according to claim 2. A learning model for causing a computer to function so as to output a probability of successful delivery of the package to the destination user for each of a plurality of time slots on the delivery date when a user feature representing a feature of a destination user of the package, a date feature representing a feature of a delivery date of the package, and a building feature representing a building type of a building located at a destination address of the package are input, the learning model comprising:
The user features include factual features of the destination user and estimated features estimated from the factual features using a learning model for machine learning;
The building type is a learning model that is identified based on the numbers and non-numeric parts contained in the string of the destination address, the position of the numbers in the string, the number of numeric digits contained in the string, the position of specified kanji characters in the string, and specified rules regarding the notation format of the numbers contained in the string, and is one of "business building" which refers to an office or company building, "unit building" which refers to a building, condominium, or apartment with multiple living spaces, "home," or "unknown." An information processing method executed by an information processing device,
a user characteristic acquisition step of acquiring a user characteristic representing a characteristic of a destination user of the package from a user device of the destination user ;
A date feature acquisition step of acquiring a date feature representing a characteristic of a delivery date of the package;
A step of identifying the building type of the building located at the destination address of the package as one of a "business building" which refers to a business or company building, a "unit building" which refers to a building, condominium, or apartment having multiple living spaces, a "house," or an "unknown type" ;
a prediction step of predicting a probability of successful delivery of the package to the destination user for each of a plurality of time slots on the delivery date based on the user characteristics, the date characteristics, and the building type;
having
The user features include factual features of the destination user and estimated features estimated from the factual features using a first learning model for machine learning;
This information processing method is characterized in that in the identification process, the building type is identified based on predetermined rules regarding one or more specified characters contained in the string of the destination address, the position of the one or more characters in the string, the position and number of digits of the numbers contained in the string, and the notation format of the numbers contained in the string . An information processing program for causing a computer to execute information processing, the program comprising:
a user characteristic acquisition process for acquiring a user characteristic representing a characteristic of a destination user of the package from a user device of the destination user ;
A date feature acquisition process for acquiring date features representing the characteristics of the delivery date of the package;
A process of identifying the building type of the building located at the destination address of the package as one of "business building" which refers to a business or company building, "unit building" which refers to a building, condominium, or apartment having multiple living spaces, "house," or "unknown type" ;
a prediction process for predicting a probability of successful delivery of the package to the destination user for each of a plurality of time slots on the delivery date based on the user characteristics, the date characteristics, and the building type ,
The user features include factual features of the destination user and estimated features estimated from the factual features using a first learning model for machine learning;
the identification process includes identifying the building type based on predetermined rules regarding one or more predetermined characters included in a character string of the destination address, the positions of the one or more characters in the character string, the positions and number of digits of numbers included in the character string, and a notation format of the numbers included in the character string;
Information processing program.

Images