JP7136730B2 - Product recommendation device and information processing program - Google Patents

Description

The embodiments of the present invention relate to a product recommendation device and an information processing program.

Various mechanisms have been considered for recommending products that match the preferences set by the purchaser to the purchaser.
For example, for "curry roux" as a product category, five index values from 1 to 5 are set for the preference item "spiciness", and the index value that the purchaser thinks matches his/her taste. set. In addition, for each product belonging to "curry roux", the seller such as the manufacturer or the store determines the index value of the "spiciness" of the product. Then, when a product related to "curry roux" should be recommended to the purchaser, it is conceivable that the product having an index value set closer to the index value determined by the purchaser is determined as the recommended product.

However, the setting of the index value by the purchaser does not necessarily match the index value determined for the product, depending on the subjectivity of the purchaser. Therefore, even if the product determined as described above is recommended to the purchaser, the product may not match the taste of the purchaser.
Under these circumstances, it has been desired to recommend products that match the preferences of the purchaser.

JP 2004-264557 A

A problem to be solved by the present invention is to provide a product recommendation device and an information processing program capable of recommending products that match the tastes of purchasers.

A product recommendation device of an embodiment includes acquisition means, determination means, recommendation means, setting means, and change means. The acquisition means acquires a preference index value determined by the purchaser, which is an index value related to the preference of the product category to be recommended. The determination means selects, as recommended products, products belonging to the product category and having a predetermined relationship between the product index value determined in advance as the index value related to the preference of the product and the preference index value acquired by the acquisition unit. decide. The recommending means recommends the recommended merchandise determined by the determining means to the purchaser. The setting means is such that the recommended product is purchased by the purchaser after being recommended by the recommending means, and after that, a product index value different from the product index value related to the recommended product is set, and the recommended product is different from the recommended product belonging to the product category. When one or a plurality of separate products are continuously purchased by the purchaser, an index value related to product category preferences is set as an evaluation index value based on the product index values determined for the separate products. The changing means changes the preference index value so that the difference between the evaluation index value set by the setting means and the preference index value becomes smaller.

FIG. 2 is a block diagram showing the main circuit configuration of a receipt server according to one embodiment; FIG. 2 is a diagram schematically showing the structure of a product record included in the sales side index table in FIG. 1; FIG. 2 is a diagram schematically showing the structure of a preference record included in the purchase side index table in FIG. 1; FIG. 2 is a diagram schematically showing the structure of a management record included in a management table in FIG. 1; FIG. FIG. 2 is a flow chart of recommended processing by the processor in FIG. 1; FIG. 6 is a diagram showing the configuration of a management record added to a management table in ACT6 in FIG. 5; FIG. FIG. 2 is a flowchart of management processing by a processor in FIG. 1; FIG. FIG. 8 is a diagram showing the configuration of the management record after being updated in ACT 13 in FIG. 7; FIG. 8 is a diagram showing the configuration of the management record after being updated in ACT 16 in FIG. 7; FIG. 8 is a diagram showing the configuration of a management record after being updated in ACT 18 in FIG. 7;

An example of an embodiment will be described below with reference to the drawings. In this embodiment, a receipt server having a function as a product recommendation device will be described as an example.
FIG. 1 is a block diagram showing the essential circuitry of a receipt server 1 according to this embodiment.

The receipt server 1 includes a processor 11, a main memory 12, an auxiliary storage device 13, a communication interface 14, a transmission line 15, and the like. The processor 11 , main memory 12 , auxiliary storage device 13 and communication interface 14 are connected by a transmission line 15 .
In the receipt server 1, the processor 11, the main memory 12, and the auxiliary storage device 13 are connected via the transmission line 15 to form a computer that performs information processing for controlling the receipt server 1. FIG.

The processor 11 corresponds to the central portion of the computer. The processor 11 controls each part to realize various functions of the receipt server 1 according to information processing programs such as an operating system, middleware, and application programs.

The main memory 12 corresponds to the main memory portion of the computer. Main memory 12 includes a non-volatile memory area and a volatile memory area. The main memory 12 stores the above information processing program in a nonvolatile memory area. The main memory 12 may also store data necessary for the processor 11 to execute processing for controlling each unit in a non-volatile or volatile memory area. The main memory 12 uses a volatile memory area as a work area in which data is appropriately rewritten by the processor 11 .

The auxiliary storage device 13 corresponds to the auxiliary storage portion of the computer. As the auxiliary storage device 13, for example, an EEPROM (electric erasable programmable read-only memory), HDD (hard disc drive) or SSD (solid state drive) can be used singly or in combination. The auxiliary storage device 13 stores data used by the processor 11 in performing various processes and data generated by the processes performed by the processor 11 . The auxiliary storage device 13 may store the above information processing program. In this embodiment, the auxiliary storage device 13 includes an application program (hereinafter referred to as a receipt server application) AP1 for operating as the receipt server 1 and an application program (hereinafter referred to as a recommendation application) describing information processing described later. ) stores AP2. The auxiliary storage device 13 also stores a transaction database DB1, a sales side index table TD1, a purchase side index table TD2 and a management table TD3. The transaction database DB1 is a database in which data contained in transaction data collected from the POS system 2 is accumulated. The transaction database DB1 represents at least the member code of the purchaser, the merchandise code of the merchandise purchased by the purchaser, and the date and time of purchase in association with each other. The sales side index table TD1, the purchase side index table TD2 and the management table TD3 will be described later.

The communication interface 14 is an interface for data communication with the POS system 2 and the purchaser terminal 3 via the communication network 4 . As the communication interface 14, for example, a well-known communication device for performing data communication via the Internet can be used.
The transmission line 15 includes an address bus, a data bus, a control signal line, etc., and transmits data and control signals exchanged between the connected units.

The receipt server 1 can use, for example, a general-purpose computer device as basic hardware. At this time, typically, the computer device in which the receipt server application AP1 and the recommendation application AP2 are not stored in the auxiliary storage device 13 and the receipt server application AP1 and the recommendation application AP2 are individually transferred to the operator of the receipt server 1. be. The receipt server application AP1 and recommendation application AP2 can be transferred by recording them on a removable recording medium such as a magnetic disk, magneto-optical disk, optical disk, or semiconductor memory, or by downloading them via a network. In this case, the receipt server application AP1 and the recommendation application AP2 are written to the auxiliary storage device 13 in accordance with the operation of the administrator of the receipt server 1 or the installation worker of the receipt server 1 or the like. Note that the transfer of the receipt server application AP1 and the recommendation application AP2 and the writing to the auxiliary storage device 13 may be performed separately.

FIG. 2 is a diagram schematically showing the structure of record data (hereinafter referred to as product record) RD1 included in the sales side index table TD1.
The sales side index table TD1 is table data including at least one product record RD1 structured as shown in FIG. One product record RD1 is associated with one product that can be recommended to the purchaser. Product record RD1 includes fields F11, F12, F13, and F14. The product record RD1 may include fields after field F15.

A product code for identifying the product associated with the product record RD1 is set in the field F11. A category code for identifying the product category to which the product associated with the product record RD1 belongs is set in the field F12. The name of the product category (hereinafter referred to as category name) is set in the field F13. In the field F14, one of the item names (hereinafter referred to as preference item name) relating to the preference of the product associated with the product record RD1 is set. Data set in the field F14 as preference item names are, for example, "spiciness", "hardness", or "spiciness". In the field F15, an index value (hereinafter referred to as a product index value) set for the preference item name set in the field F14 regarding the product associated with the product record RD1 is set. A single preference item name different from that of field F14 is set in field F16. The field F17 is set with the product index value set for the preference item name set in the field F16 for the product associated with the product record RD1. Thus, each field after the field F14 is set in pairs, and a preference item name and a product index value are set. The number of fields after the field F16 varies depending on the number of preference item names for which product index values are set for products associated with the product record RD1. Note that the product index value is appropriately set by the sales side, such as the manufacturer or distributor, in consideration of the properties of the product.

In this embodiment, the product category is a category in which products that can be replaced with each other, such as “curry roux,” are grouped together. For example, sweet curry roux made by Company A, medium spicy curry roux made by Company A, dry curry roux made by Company A, sweet curry roux made by Company B, and dry curry roux made by Company B are called “curry roux”. The product belongs to the product category.

FIG. 3 is a diagram schematically showing the structure of record data (hereinafter referred to as preference record) RD2 included in the purchase side index table TD2.
The purchase side index table TD2 is table data containing at least one preference record RD2 structured as shown in FIG. One preference record RD2 is associated with a combination of an electronic receipt service member and a product category. Preference record RD2 includes fields F21, F22, F23, F24 and F25. The preference record RD2 may include fields after field F26.

A member code for identifying a member associated with the preference record RD2 is set in the field F21. The field F22 is set with the category code of the product category associated with the preference record RD2. The field F23 is set with the category name of the product category associated with the preference record RD2. One of the preference item names related to the product category associated with the preference record RD2 is set in the field F24. In the field F25, an index value (hereinafter referred to as a preference index value) set for the preference item name set in the field F24 for the product category associated with the preference record RD2 is set. A single preference item name different from that of field F24 is set in field F26. The field F27 is set with the product index value set for the preference item name set in the field F26 for the product category associated with the preference record RD2. Thus, each field after the field F24 is a set of two, and a preference item name and a preference index value are set. The number of fields after the field F26 varies depending on the number of preference item names for which preference index values are set for the product category associated with the preference record RD2. The preference index value is appropriately set by the member according to the member's own preference. The preference index value may be corrected as described later. The preference index value may be reset by the member.

FIG. 4 is a diagram schematically showing the structure of record data (hereinafter referred to as a management record) RD3 included in the management table TD3.
The management table TD3 is table data containing at least one management record RD3 structured as shown in FIG. One management record RD3 is associated with a combination of a member and a product recommended to the member. Management record RD3 includes fields F31, F32, F33, and F34. The management record RD3 may include fields after field F35. Various data are set in each field of the management record RD3 by the processor 11 through information processing to be described later. Data set in each field of the management record RD3 will be described later.

Next, the operation of the receipt server 1 configured as above will be described. It should be noted that the contents of the various processes described below are only examples, and it is possible to change the order of some of the processes, omit some of the processes, or add other processes as appropriate.

A feature of the operation of the receipt server 1 in this embodiment is the operation for referring to the collected transaction data and recommending products to the purchaser while considering the purchase history of the purchaser. The information processing executed by the processor 11 according to, for example, the receipt server application AP1, that is, the information processing for collecting transaction data from the POS system 2 and managing the transaction database DB1, is performed by, for example, an existing electronic receipt system. You can use the motion that is Therefore, in the following description, descriptions of operations that can utilize these existing operations will be omitted. The transaction data is data related to the purchaser's purchase of products from the seller. and at least.

At a predetermined recommended timing, the processor 11 executes the following information processing (hereinafter referred to as recommendation processing) in accordance with the recommendation application AP2. The recommended timing is, for example, the timing at which a pre-scheduled date and time has arrived. The recommended timing is, for example, the timing when it is detected that the administrator of the receipt server 1 or the like has given a start instruction. The recommendation timing is, for example, the timing at which it is detected that the purchaser terminal 3 has requested the recommendation of a product. The recommended timing is, for example, the timing at which it is detected that the purchaser has performed a predetermined action, such as when the purchaser purchases a specific product or when the purchaser enters a store. The timing to be the recommended timing may be appropriately determined by the designer of the receipt server 1, the creator of the recommendation application AP2, the administrator of the receipt server 1, or the like.

FIG. 5 is a flow chart of recommendation processing by the processor 11 .
As ACT 1, the processor 11 determines a product category to be recommended as a target category. For example, the processor 11 sets, as a target category, a product category scheduled in advance for the recommendation timing at which the recommendation process is started. The processor 11, for example, sets the product category specified by the manager or the purchaser as the target category. For example, the processor 11 selects a product category selected according to a predetermined rule or randomly from a list of predetermined product categories as a target category. How to determine the target category may be appropriately determined by the above designer, creator, or administrator.

As ACT2, the processor 11 determines purchasers to be recommended (hereinafter referred to as target purchasers) from among the members. For example, the processor 11 selects members scheduled in advance for the start timing as target purchasers. For example, the processor 11 selects members designated by the administrator as target purchasers. The processor 11, for example, selects the member who requested the recommendation as the target purchaser. For example, the processor 11 selects a member selected according to a predetermined rule or randomly from a predetermined list of members as the target purchaser. How to determine the target purchasers may be determined as appropriate by the designer, creator, or manager described above.

As ACT3, the processor 11 extracts the preference record RD2 associated with the combination of the determined target category and the member who is the purchaser from the purchase side index table TD2. For example, the processor 11 stores the preference record RD2 in which the category code of the target category determined in ACT1 is set in the field F22 and the membership code of the target purchaser determined in ACT2 is set in the field F21. Extract from TD2. As a result, the processor 11 acquires the preference index value determined by the purchaser, which is an index value related to the preference of the product category to be recommended. As a result of the processor 11 executing information processing based on the recommendation application AP2, the computer having the processor 11 as its central part functions as an acquisition means.

As ACT 4, the processor 11 determines the product to be recommended by the current recommendation process (hereinafter referred to as the recommended product). Here, the processor 11 refers to the data set in each field after the field F24 of the preference record RD2 extracted in ACT3. The processor 11 also refers to the data set in each field after the field F14 of the product record RD1 associated with each product belonging to the target category. Then, the processor 11 determines recommended products according to a predetermined algorithm. As a result of the processor 11 executing information processing based on the recommendation application AP2, the computer having the processor 11 as its central part functions as determining means for determining recommended products.

For example, the processor 11 narrows down one product record RD1 under the following conditions. The processor 11 then determines the product identified by the product code set in the field F11 of the product record RD1 as the recommended product.
(1) The category code of the target category is set in field F12.
(2) The same preference item name as the preference item name set in each field after field F24 of the preference record RD2 extracted in ACT3 (hereinafter referred to as a common item name) is in one of the fields after field F14. is set. In addition, the product index value set in the field next to the field in which the common item name is set in the product record RD1 is the preference set in the field next to the field in which the common item name is set in the preference record RD2. match the index value.
(3) The number of common item names whose product index values match the preference index values is greater.
(4) The product index value set in the field next to the field in which the common item name is set in the product record RD1 is set in the field next to the field in which the common item name is set in the preference record RD2. Fewer common items that differ from the preference index value.
(5) fewer preference item names separate from common item names;
(6) Random selection.

Note that other conditions such as the priority set for each product record RD1 or the importance set for each preference item name may be added to the conditions for narrowing down.
Also, instead of the above (1) to (5), an algorithm for determining recommended products, such as applying "the total difference between the product index value and the preference index value for each common item name is smaller." may be appropriately determined by the designer of the receipt server 1, the creator of the recommendation application AP2, the manager of the receipt server 1, or the like.

As a specific example, as the preference index value corresponding to the preference item name “spiciness” for the product category whose category name is “curry roux”, “5” out of five index values from “1” to “5” is selected. It is assumed that the purchaser has set In this case, the processor 11 sets "5" as the product index value corresponding to the preference item name "spiciness" among the products belonging to the product category whose category name is "curry roux" for this purchaser. One of the products in the list is set as the recommended product.
However, the processor 11 may determine a plurality of commodities that match the conditions as recommended commodities.

In ACT 5, the processor 11 executes notification processing for notifying the target purchaser of the recommended product. This notification process is, for example, a process of sending an e-mail appealing the recommended product to the target purchaser to an e-mail address associated with the target purchaser in advance. The notification process is, for example, a so-called push notification process that requests the purchaser terminal 3 pre-associated with the target purchaser to display a notification screen that appeals the recommended product to the target purchaser. and Alternatively, the notification process may be, for example, a process for displaying on the purchaser terminal 3 a web page that appeals the recommended product to the target purchaser. Recommended products are recommended to the purchaser by this notification process. Thus, the processor 11 executes information processing based on the recommendation application AP2, so that the computer having the processor 11 as a central part functions as recommendation means.

As ACT6, the processor 11 updates the management table DT3. That is, the processor 11 adds to the management table DT3 the management record RD3 associated with the combination of the target purchaser and the recommended product in the product recommendation made this time.

FIG. 6 shows the configuration of the management record RD3 added to the management table DT3 in ACT6. Note that FIG. 6 is one of several drawings showing the transition of one management record RD3. However, in order to facilitate discrimination of the management records RD3 in different states shown in these drawings, the reference numeral is designated as "RD3-1" for convenience.
In the management record RD3-1 shown in FIG. 6, the member code of the target purchaser is "0023561", the product code of the recommended product is "4901234567890", and the notification processing is at 10:00 on January 1, 2019. This is an example of a case where

Therefore, in ACT6, the processor 11 sets "0023561", "4901234567890" and "2019/01/01 10:00" in fields F31, F32 and F33, respectively, and resets the purchase flag as shown in FIG. A management record RD3-1 is generated as shown. The processor 11 then updates the management table DT3 to newly include this management record RD3-1.
After updating the management table DT3, the processor 11 ends the recommendation process.

On the other hand, at a predetermined management timing, the processor 11 executes information processing (hereinafter referred to as management processing) described below according to the recommendation application AP2. For example, the processor 11 repeatedly performs the management process on the management records DR3 included in the management table DT3. In this case, the management timing is the timing at which the processor 11 finishes management processing for one management record DR3 and can start management processing for another management record DR3. Note that the processor 11 may execute a plurality of management processes in parallel. The timing to be used as the management timing may be appropriately determined by the designer of the receipt server 1, the creator of the recommendation application AP2, the administrator of the receipt server 1, or the like.

FIG. 7 is a flow chart of management processing by the processor 11 . Incidentally, in the following description regarding FIG. 7, simply describing "management record RD3" indicates the management record RD3 that is the object of management processing.
As ACT11, the processor 11 confirms whether or not the purchase flag set in the field F34 of the management record RD3 is set. Then, if the purchase flag is in a reset state, the processor 11 determines NO, and proceeds to ACT12.

As ACT12, the processor 11 confirms whether or not the recommended product associated with the management record RD3 has been purchased by the target purchaser after the recommendation. For example, the processor 11 associates the merchandise code set in the field F32 of the management record RD3 with the membership code set in the field F31 of the management record RD3, and stores the merchandise code in the transaction database DB1, and the purchase date and time of the merchandise code in the management record RD3. If it is later than the date and time of implementation set in the field F33, it is determined as YES, and the process proceeds to ACT13.

As ACT13, the processor 11 updates the management record RD3. That is, the processor 11, for example, changes the purchase flag set in the field F34 of the management record RD3 to the set state. Also, the processor 11, for example, adds a field F35 to the management record RD3, and sets the purchase date and time of the recommended product shown in the transaction database DB1 in the field F35.

FIG. 8 is a diagram showing the structure of the management record RD3 after being updated in ACT13. Note that FIG. 8 shows an update example in the case where the management record RD3-1 shown in FIG. 6 is the management record before update, and is denoted by "RD3-2" for the sake of convenience.
The management record RD3-2 shown in FIG. 8 is an example when the recommended product is purchased at 17:23 on January 5, 2019.
For this reason, in ACT13, the processor 11 changes the purchase flag to the set state, and adds the field F35 in which "2019/01/05 17:23" is set to the management record RD3-1 as the management record RD3-2. there is

After updating the management record RD3 in ACT13, the processor 11 terminates the management process. Note that the processor 11, for example, when the product code set in the field F32 of the management record RD3 is not included in the transaction database DB1 in association with the member code set in the field F31 of the management record RD3, or when such a product Although the code is included in the transaction database DB1, if the purchase date and time is earlier than the implementation date and time set in the field F33 of the management record RD3, NO is determined in ACT12, and ACT13 is not performed. End management processing.
That is, the processor 11 waits for the purchase of the recommended product by the corresponding purchaser by repeating the management processing for the same management record RD3. Then, the processor 11 sets the purchase flag when the recommended product is purchased by the corresponding purchaser.

If the purchase flag set in the field F34 of the management record RD3 is set when the management process is started, the processor 11 determines YES in ACT11 and proceeds to ACT14. That is, processor 11 advances to ACT 14 if the recommended product has already been purchased.
As ACT 14, the processor 11 confirms whether or not the purchase of the substitute product has been detected. A substitute product is a product that belongs to the same product category as the recommended product, is different from the recommended product, and is purchased after the recommended product. For example, when the management record RD3 does not include the field F36 as in the management record RD3-2 shown in FIG. 8, the processor 11 determines NO and proceeds to ACT15.

As ACT 15, the processor 11 confirms whether or not the substitute product has been purchased. For example, the processor 11 determines a product that meets the following conditions as a substitute product. If there are a plurality of commodities that match the conditions, the processor 11 determines all of them as substitute commodities.
(1) The product belongs to the same product category as the product identified by the product code set in the field F32 of the management record RD3, and is identified by a product code different from the above product code.
(2) It is identified by the product code contained in the transaction database DB1 in association with the member code set in the field F31 of the management record RD3.
(3) The purchase date and time shown in the transaction database DB1 is later than the purchase date and time set in the field F35 of the management record RD3.
If the processor 11 can determine even one substitute product, it determines YES, and proceeds to ACT16.

As ACT16, the processor 11 updates the management record RD3. In other words, the processor 11 updates the management record RD3 so as to add two fields at the end, in which the product code for identifying the determined alternative product and the date and time of purchase thereof are respectively set. If a plurality of substitute products are determined in ACT 15, the processor 11 sets a set of two fields in which the product code for identifying the substitute product and the purchase date and time are set, respectively. Update the management record RD3 so as to add the number of products.

FIG. 9 is a diagram showing the configuration of the management record RD3 after being updated in ACT16. Note that FIG. 9 shows an update example in the case where the management record RD3-2 shown in FIG. 8 is the management record before update, and the code is "RD3-3" for convenience.
The management record RD3-3 shown in FIG. 9 is an example of a case where a substitute product whose product code is "4909999999999" was purchased at 18:37 on January 21, 2019.
Therefore, in ACT16, the processor 11 adds a field F36 in which "4909999999999" is set and a field F37 in which "2019/01/21 18:37" is set to the management record RD3-2. and

After updating the management record RD3 in ACT16, the processor 11 terminates the management process. If the processor 11 cannot confirm that the substitute product has been purchased, it determines NO in ACT15 and terminates the management process without performing ACT16.
In other words, the processor 11 waits for the purchase of a substitute product by repeating the management process in a situation where the recommended product has already been purchased but no substitute product has been purchased. Then, when the substitute product is purchased, the processor 11 updates the management record RD3 so as to represent the purchase history.

When the management process is started, the processor 11 determines YES in ACT14 if the management record RD3 includes the field F35, and proceeds to ACT17. That is, the processor 11 advances to ACT 17 if the purchase of the substitute item has been confirmed in the past.

In ACT 17, the processor 11 confirms whether or not a substitute product other than the substitute product whose purchase has been confirmed has been newly purchased. For example, the processor 11 determines a product that meets the following conditions as a newly purchased substitute product. If there are a plurality of commodities that match the conditions, the processor 11 judges them as newly purchased substitute commodities.
(1) The product belongs to the product category to which the product identified by the product code set in the field F32 of the management record RD3 belongs and is identified by the product code different from the recommended product.
(2) It is identified by the product code contained in the transaction database DB1 in association with the member code set in the field F31 of the management record RD3.
(3) The date and time of purchase is later than the date and time of purchase set in the last field of the management record RD3.
Then, if the processor 11 cannot determine even one newly purchased substitute product, it determines NO, and ends the management process as it is. In other words, in a situation where a substitute product is not newly purchased, the processor 11 waits for a substitute product to be newly purchased by repeating the management process. Then, if the processor 11 can confirm that the purchased product has been newly purchased, it determines YES in ACT17 and proceeds to ACT18.

As ACT18, the processor 11 updates the management record RD3. In other words, the processor 11 updates the management record RD3 so as to add two fields at the end, in which the commodity code for identifying the purchased substitute commodity and the purchase date and time are respectively set. If the processor 11 determines a plurality of substitute products in ACT 17, the processor 11 sets a set of two fields in which the product code for identifying the substitute product and the date and time of purchase thereof are respectively set. Update the management record RD3 so as to add the number of products.

FIG. 10 is a diagram showing the structure of the management record RD3 after being updated in ACT18. Note that FIG. 10 shows an update example in the case where the management record RD3-3 shown in FIG. 9 is the management record before update, and is denoted by "RD3-4" for the sake of convenience.
The management record RD3-4 shown in FIG. 10 is an example of a case in which a substitute product whose product code is "4902233445566" was purchased on January 27, 2019 at 15:49.
Therefore, in ACT18, the processor 11 adds a field F38 in which "4902233445566" is set and a field F39 in which "2019/01/27 15:49" is set to the management record RD3-3. and
In this way, the management record RD3 records a history of purchase of substitute products. When the processor 11 advances to ACT 18, it is a case where the substitute product is purchased continuously over a plurality of times after the recommended product is purchased.

As ACT 19, the processor 11 determines the evaluation index value. The evaluation index value is an index value of the purchaser's preference evaluated based on the product index value set for the alternative product. For example, the processor 11 extracts all product records RD1 associated with substitute products from the seller index table TD1. The product record RD1 to be extracted will be referred to as substitute product record RD1 here for convenience. Further, the processor 11 sets a member code matching the member code set in the field F31 of the management record RD3 in the field F21, and sets the category code of the product category to which the recommended product associated with the management record RD3 belongs to the field F22. The set preference record RD2 is selected from the purchase side index table TD2. The selected preference record RD2 is here called a correction target record RD2 for convenience. The processor 11 determines an evaluation index value for each preference item name set in the correction target record RD2 based on the product index value set in the substitute item record RD1 in association with the preference item name. Although there are at least two replacement item records RD1, the preference item name set in the correction target record RD2 is not necessarily set in each of the replacement item records RD1. In other words, the number of product index values to be used for evaluation index value determination is indefinite. However, it is desirable to create the sales side index table TD1 and the purchase side index table TD2 so that the preference item names set in the correction target record RD2 are set in each substitute item record RD1 as much as possible.

An algorithm for the processor 11 to determine the evaluation index value may be appropriately determined by the designer of the receipt server 1, the creator of the recommendation application AP2, or the administrator of the receipt server 1, or the like. As an example, if there is no product index value to be used for evaluation index value determination, the same value as the preference index value set in association with the same preference item name in the correction target record RD2 is used as the evaluation index value. . Alternatively, if there is no product index value to be used for determining the evaluation index value, the average value of the values that can be set as the preference index value may be used as the evaluation index value. If there are one or more product index values to be used for determining the evaluation index value, the average value of those product index values is used as the evaluation index value. When there are one or more product index values to be used for determining the evaluation index value, a statistical value other than the average value, such as the maximum value of those product index values, may be used as the evaluation index value.

As a specific example, among the products belonging to the product category whose category name is “curry roux,” one of the products for which “5” is set as the product index value corresponding to the preference item name “spiciness” is selected. In the case of a recommended product, although the recommended product was purchased once, there are products for which "3" is set as the product index value corresponding to the preference item name "spiciness" twice after that. Suppose it is purchased. In this case, the processor 11 sets the evaluation index value to "3", for example.
As a result of the processor 11 executing information processing based on the recommendation application AP2, the computer having the processor 11 as a central part functions as setting means for setting evaluation index values.

As ACT20, the processor 11 corrects the purchase side index value set in the correction target record RD2. An algorithm for the processor 11 to correct the purchaser side index value may be appropriately determined by the designer of the receipt server 1, the creator of the recommendation application AP2, the manager of the receipt server 1, or the like. As an example, the target purchase side index value is replaced with the evaluation index value determined in ACT 19 for the same preference item name. Alternatively, the purchase side index value of interest and the evaluation index value determined in ACT 19 for the same preference item name are compared, and if the former is lower than the latter, the purchase side index value of interest is increased by a certain value, and the former is the latter. If it is higher than , the purchase side index value of the target is decreased by a certain value. The algorithm is defined so as to reduce at least the difference between the target purchaser side index value and the evaluation index value determined in ACT 19 with respect to the same preference item name.
As a specific example, if the purchase side index value corresponding to the preference item name "spiciness" is "5" and the evaluation index value is "3" as described above, the processor 11 sets the purchase side index value to " 3”. Alternatively, the processor 11 updates the purchase index value to "4" by subtracting a constant value "1" from "5" before correction.

As a result of the processor 11 executing information processing based on the recommendation application, the computer having the processor 11 as its central part functions as changing means for changing the preference index value so as to reduce the difference between the evaluation index value and the preference index value. do.
Then, when processor 11 finishes updating the purchase side index value in ACT 20, processor 11 ends the management process.

As described above, according to the receipt server 1, when the purchaser continues to purchase another substitute product after purchasing the recommended product, the product index value related to the product purchased as the substitute product is considered. The preference index value is corrected so as to reduce the difference between the determined evaluation index value and the preference index value for the purchaser. As a result, if the recommended product does not match the purchaser's taste because the preference index value set by the purchaser is not appropriate, and the purchaser purchases a substitute product for that reason, the preference index value will be reduced. will be corrected. Then, the recommended products determined based on the corrected preference index values are more likely to be products that match the preferences of the purchaser. In other words, it becomes possible to recommend products that match the preferences of the purchaser.

Moreover, according to this embodiment, the preference index value is not corrected in a situation where the recommended product has never been purchased. Therefore, in a situation where the purchaser has not yet tried the recommended product, the product is repeatedly recommended while respecting the set preference index value. In other words, the purchaser's settings are respected to some extent, rather than simply recommending products based on purchase records.

Further, according to the receipt server 1, each time a substitute product is newly purchased, the purchaser side index value is corrected in consideration of the product index value of each substitute product purchased in the past. For this reason, as the number of substitute products increases, the recommended products are selected by reflecting the purchaser's preferences, which are determined based on the purchase history.

This embodiment can be modified in various ways as follows.
It is also possible to realize a product recommendation device that does not have a function as a receipt server and is configured to specifically execute each of the above processes by referring to the transaction database DB1 managed by another receipt server. . Alternatively, the function of executing each of the above processes may be provided as one function of another device such as a POS server.

The product index value referred to for determining the product index value may be limited to substitute products purchased within a certain recent period. In this way, it is possible to recommend products that match the recent tastes of the purchaser when the tastes of the purchaser change.

In ACT12, the processor 11 may determine YES not only the recommended product itself but also a product equivalent to the product identified by the product code has been purchased. In other words, when considering the purchase of a recommended product, the purchaser may compare and examine other products of the same type displayed near the recommended product, and purchase another product based on the judgment that it is equivalent. could be. Therefore, in such a case, if the substitute product is continuously purchased after that, it can be considered equivalent to the case where the recommended product does not suit the purchaser's taste.

Each function realized by the processor 11 by information processing can be partially or entirely realized by hardware such as a logic circuit that executes information processing not based on a program. Further, each of the functions described above can be implemented by combining hardware such as the logic circuit described above with software control.

While several embodiments of the invention have been described, these embodiments have been presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the scope of the invention described in the claims and equivalents thereof.

DESCRIPTION OF SYMBOLS 1... Receipt server, 2... POS system, 3... Buyer terminal, 4... Communication network, 11... Processor, 12... Main memory, 13... Auxiliary storage device, 14... Communication interface, 15... Transmission path.

Claims (2)

Acquisition means for acquiring a preference index value determined by a purchaser, which is an index value related to the preference of a product category to be recommended;
A product belonging to the product category and having a predetermined relationship with the preference index value obtained by the obtaining means is determined as a recommended product. a determining means to
a recommendation means for recommending the recommended product determined by the determination means to the purchaser;
The recommended product is purchased by the purchaser after being recommended by the recommendation means, and thereafter, the product index value different from the product index value related to the recommended product is determined, and the recommended product belongs to the product category. When one or more separate products different from the product are continuously purchased by the purchaser, the index value related to the preference of the product category is evaluated based on the product index value set for the separate product. setting means for setting as an index value;
changing means for changing the preference index value so as to reduce the difference between the evaluation index value set by the setting means and the preference index value;
A product recommendation device equipped with the computer,
Acquisition means for acquiring a preference index value determined by a purchaser, which is an index value related to the preference of a product category to be recommended;
A product belonging to the product category and having a predetermined relationship with the preference index value obtained by the obtaining means is determined as a recommended product. a determining means to
a recommendation means for recommending the recommended product determined by the determination means to the purchaser;
The recommended product is purchased by the purchaser after being recommended by the recommendation means, and thereafter, the product index value different from the product index value related to the recommended product is determined, and the recommended product belongs to the product category. When one or more separate products different from the product are continuously purchased by the purchaser, the index value related to the preference of the product category is evaluated based on the product index value set for the separate product. setting means for setting as an index value;
changing means for changing the preference index value so as to reduce the difference between the evaluation index value set by the setting means and the preference index value;
Information processing program to function as

Images