JP2012203454A - Recommendation list generator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a recommendation list generator which generates a recommendation list for recommending proper commodities for unessential purchase, which are to be purchased in combination with a main commodity, to a user.SOLUTION: A recommendation list generator 11 includes: unessential purchase candidate commodity group extraction means 112 which extracts a combination of a plurality of commodities associated with each other with a prescribed frequency or higher, as an unessential purchase candidate commodity group; co-occurrence rate calculation means 113 which searches Web pages with the commodity names of the plurality of commodities included in the unessential purchase candidate commodity group as search queries and calculates co-occurrence rates being rates of appearance of these queries in the same Web pages among all Web pages where these search queries appear; unessential purchase candidate commodity group narrowing-down means 114 which excludes a plurality of commodities the co-occurrence rates of which are equal to or lower than a prescribed value from the unessential purchase candidate commodity group; and recommendation list generating means 115 which stores the unessential purchase candidate commodity group into recommendation list storage means 118 as a recommendation list.

Description

  The present invention relates to a recommendation list generating apparatus.

  2. Description of the Related Art Conventionally, a system that sells and purchases products using a computer via a network such as the Internet has been widespread. According to such a system, a user can sell a product on a Web page or purchase a product.

When a user purchases a product on a Web page using such a system, the user first selects a product to be purchased on a product page that provides product information, and then moves to a payment page to settle the product. be able to.
In Patent Document 1, as a system that enables settlement of a product on a Web page, a request for product purchase transmitted from a user terminal is authenticated using an authentication key included in the request, and the product A merchandise sales system that performs settlement processing has been proposed.
According to this product sales system, the user can make a payment for a product by making a request for purchasing the product.

JP 2008-111231 A

  By the way, in a system that enables payment of a product on a Web page, when a user purchases a product, the purchase price of the product reaches a predetermined reference amount, thereby reducing a delivery fee or the like. May provide services. Therefore, when a user makes a request for purchase of a product and settles the product, if the purchase price of the product does not reach the predetermined reference amount for receiving the service, the product is set as the main product. It may be possible to adjust so that the purchase amount reaches a predetermined reference amount by adding a so-called purchase product to be purchased in combination with the main product.

  However, in such a case, the user moves again from the payment page to the product page, considers what kind of purchase product to add, decides the purchase product to add, and then returns to the payment page. It is necessary to perform a complicated operation of confirming whether the total purchase amount reaches a predetermined reference amount. Therefore, a user who dislikes such a complicated operation may make a settlement as it is without making a purchase. In addition, when the user who has made a purchase returns to the product page again, the purchase may be canceled as it is. Such a system makes the operation complicated for the user and at the same time misses the sales opportunity for the merchant who sells the product. Therefore, it is conceivable to perform a recommendation that allows the user to select an appropriate purchase product that reaches a predetermined reference amount on the spot without moving the user from the payment page.

  The present invention has been made in view of such a problem, and provides a recommendation list generation device for creating a recommendation list for recommending a user to purchase a product that is appropriately purchased in combination with a main product. With the goal.

  The present invention provides the following solutions.

(1) A recommendation list generation device that generates a recommendation list for transmitting recommended information of a purchased product to be purchased in combination with a main product to a user terminal connectable via a network,
Purchase log storage means for associating and storing a plurality of products purchased by the same user within a predetermined period as purchase logs;
Referring to the purchase log stored in the purchase log storage means, and then extracting a combination of the plurality of products associated with a predetermined number of times or more as a purchase candidate product group;
Next, a Web page is searched using the product names of the plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extraction unit, and among all the Web pages in which these search queries appear. A co-occurrence rate calculating means for calculating a co-occurrence rate, which is the rate at which these search queries appear on the same Web page;
Next, the candidate product group narrowing means for excluding the plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculating unit is a predetermined value or less from the buy candidate product group;
A recommendation list generation device comprising: a recommendation list generation unit that stores the candidate purchase product group as the recommendation list after the purchase candidate product group narrowing unit narrows down.

  The recommendation list generating device of (1) generates a recommendation list for transmitting recommendation information of a purchased product to be purchased in combination with the main product to a user terminal connectable via a network. The purchase log storage means stores a plurality of products purchased by the same user within a predetermined period as a purchase log in association with each other. Next, the purchase candidate product group extraction unit refers to the purchase log stored in the purchase log storage unit, and then extracts a combination of a plurality of products associated with a predetermined number of times as a purchase candidate product group. The co-occurrence rate calculating unit then searches the Web page using the product names of a plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extracting unit, and all of these search queries appear. A co-occurrence rate, which is a ratio of these search queries appearing on the same Web page in the Web page, is calculated. Next, the buying candidate product group narrowing means excludes a plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculating means is a predetermined value or less from the buying candidate product group. The recommendation list generation unit stores the candidate purchase product group as a recommendation list after the purchase candidate product group narrowing unit narrows down.

As a result, products that are purchased at the same time as the main product, the number of purchases of which is equal to or greater than the predetermined number of times are extracted, and then the candidate product group is extracted from the group of candidate products for purchase. A recommendation list for transmitting recommendation information of the purchased product is generated after the product is purchased in combination with the main product by excluding the product whose value is equal to or less than the predetermined value. That is, the candidate product group to be purchased after the recommendation list is purchased at the same time as the main product and is posted on the same Web page more often than other products. Therefore, for example, when purchasing another product in addition to the main product, the user can select an appropriate purchase product without performing a complicated operation if a candidate purchase product group is displayed next. .
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

(2) The purchase log storage means further stores the price of the product in association with the product,
The recommendation list generation unit refers to the purchase log storage unit, and then sets, as a main product, a product having the highest price among the plurality of products included in the candidate product group for purchase as described in (1). Recommendation list generator.

The recommendation list generation device of (2) refers to the purchase log storage means, and then sets the highest price among a plurality of products included in the candidate purchase product group as the main product.
Thereby, since the main product has the highest unit price among a plurality of products purchased by the same user during the predetermined period, it is highly likely that the main product was the main purchase item of the user. And it is highly possible that the product other than the main product was an accessory or consumable of the main product. Thus, by determining the main product, it is possible to generate a recommendation list that matches the user's shopping mode.
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

(3) The purchase candidate product group extraction unit then refers to the purchase log, sets a larger weighting value as the number of times the plurality of products are associated, and stores the weighting value in association with the purchase candidate product group. And
The recommendation list generating means stores the weight value associated with the purchase candidate commodity group by the purchase candidate commodity group extraction means and then further associated with the purchase commodity group in the recommendation list (1) or ( The recommendation list generating device according to 2).

The recommendation list generation device of (3) refers to the purchase log, sets a larger weighting value as the number of times that the plurality of products are associated, and then stores the weighting value in association with the buying candidate product group. Then, in the recommendation list, it is stored in further association with the purchased product group.
Thereby, since it is possible to rank the purchase candidate product group in the recommendation list by the weighting value, for example, it is possible to show the user in order from the purchase product having the higher weighting value.
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

  ADVANTAGE OF THE INVENTION According to this invention, the recommendation list production | generation apparatus which produces the recommendation list | wrist for recommending the next purchase goods suitable for a user purchased in combination with a main product can be provided. By utilizing this recommendation list, it is possible to make a recommendation that allows the user to select an appropriate purchase product that reaches a predetermined reference amount on the spot without moving the user from the payment page.

It is a figure which shows the function structure of the recommendation system which concerns on embodiment of this invention. It is a figure explaining purchase log DB concerning the above-mentioned embodiment. It is a figure explaining the process which calculates the weighting value by the purchase candidate goods group extraction means concerning the said embodiment. It is a figure explaining the process which calculates the co-occurrence rate by the co-occurrence rate calculation means which concerns on the said embodiment. It is a figure explaining the recommendation list | wrist which concerns on the said embodiment. It is a flowchart of the process which the recommendation list production | generation apparatus of the recommendation apparatus which concerns on the said embodiment produces | generates a recommendation list. It is a flowchart of the process in which the purchase goods recommendation apparatus next to the recommendation apparatus which concerns on the said embodiment transmits the recommendation information of purchase goods next to a user terminal.

  Embodiments of the present invention will be described below in detail with reference to the drawings. In the following description of the embodiments, the same constituent elements are denoted by the same reference numerals, and the description thereof is omitted or simplified.

[Functional configuration of recommendation system 1]
FIG. 1 is a diagram showing a functional configuration of a recommendation system 1 according to the embodiment of the present invention.
The recommendation system 1 is a system that sells and purchases products via the network 7, and includes the recommendation device 10, a plurality of user terminals 20 that can be connected to the recommendation device 10 via the network 7, and the network 7. And a search server 30 connectable to the recommendation device 10.

  The recommendation device 10 is configured to generate a recommendation list for generating recommendation list for transmitting recommendation information of a purchased product to be purchased in combination with the main product to the user terminal 20, and to the user terminal 20 based on the recommendation list. And a purchased product recommendation device 12 for notifying the recommended product. The recommendation device 10 can be realized by using various general-purpose computers and the like, and the number thereof is not limited, and the recommendation list generation device 11, then the purchased product recommendation device 12 and the search server 30 are integrated into one device. It can also be configured.

  The recommendation list generation device 11 includes a purchase log acquisition unit 111, a purchase candidate product group extraction unit 112, a co-occurrence rate calculation unit 113, a purchase candidate product group narrowing unit 114, a recommendation list generation unit 115, and a purchase. Log storage means 117 and recommendation list storage means 118 are provided.

The purchase log acquisition unit 111 acquires a purchase log, which is information about a product purchased by the user, from the user terminal 20 and stores it in a purchase log database (DB) (see FIG. 2) stored in the purchase log storage unit 117. To do.
The purchase log storage unit 117 stores a plurality of products purchased by the same user within a predetermined period in the purchase log DB as purchase logs. Further, the purchase log storage unit 117 associates the price of the product with the product and further stores it in the purchase log DB.

The purchase log database (DB) will be described with reference to FIG.
FIG. 2 is a diagram for explaining the purchase log DB.
In the purchase log DB, each purchaser ID is associated with a purchase product ID, a purchase product name, and a unit price that identify a plurality of purchase products purchased within a predetermined period.
The purchaser ID is an identifier unique to each user for identifying the user who has accessed through the user terminal 20.
The purchased product ID is an identifier unique to each product for identifying the product. In the purchase log DB, a purchase product ID of a purchase product purchased in a predetermined period (for example, at the same timing or within 24 hours) by a user having the purchaser ID is associated with each purchaser ID.
The purchased product name is the name of the type of purchased product. In the purchase log DB, each purchase product ID is associated with a purchase trade name having the purchase product ID.
The unit price is the unit price of the purchased product. In the purchase log DB, the unit price of the purchased product having the purchased product ID is associated with each purchased product ID.

Returning to FIG. 1, the purchase candidate product group extraction unit 112 then refers to the purchase log DB (see FIG. 2) stored in the purchase log storage unit 117, and a combination of a plurality of products associated with a predetermined number of times or more. Are extracted as a candidate product group for purchase.
For example, referring to FIG. 2, the purchase candidate product group extraction unit 112 then selects “notebook computer”, “optical drive”, and “ Next, a combination of “notebook computer” and “mouse” is extracted as a candidate product group.

  The purchase candidate product group extraction unit 112 then refers to the purchase log stored in the purchase log storage unit 117, sets a larger weighting value as the number of times a plurality of products are associated, and then purchase candidate products. Store in association with the group.

FIG. 3 is a diagram for explaining a process of setting a weighting value by the buying candidate commodity group extracting unit 112.
Next, with reference to the purchase log DB (see FIG. 2), the purchase candidate product group extraction unit 112 refers to the “notebook computer” and “optical drive” that are associated with the purchase product a predetermined number of times (for example, twice) or more. The combination of “notebook computer” and “mouse” is then extracted as a candidate product group. Then, the candidate product group extraction unit 112 weights each of the combinations of “notebook computer” and “optical drive” and “notebook computer” and “mouse” according to the number included in the purchase log DB. Is calculated. Specifically, the purchase candidate product group extraction unit 112 calculates the weighting value 2.0 because the combination of “notebook computer” and “optical drive” is included twice in the purchase log DB, and “notebook” Since the combination of “personal computer” and “mouse” is included three times in the purchase log DB, a weight value of 3.0 is calculated. The weighting value can also be calculated by multiplying the number of times included in the purchase log DB by a predetermined multiplier.

  Returning to FIG. 1, the co-occurrence rate calculating unit 113 then searches the Web page using the product names of a plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extracting unit 112 as search queries. The co-occurrence rate, which is the ratio of these search queries appearing on the same web page, is calculated among all the web pages in which the search query appears. Specifically, the co-occurrence rate calculating unit 113 then transmits the product names of a plurality of products included in the candidate product group to the search server 30 as a search query, and all Web pages in which these search queries appear. The number of searches is received, and the co-occurrence rate is calculated for each candidate product based on the received number of searches.

FIG. 4 is a diagram for explaining the process of calculating the co-occurrence rate by the co-occurrence rate calculating unit 113.
The co-occurrence rate calculating means 113 searches for the product names “notebook computer” and “optical drive” and “notebook computer” and “mouse”, and combinations of product names “notebook computer AND optical drive” and “notebook computer AND mouse”. It transmits to the search server 30 as a query, and each search number is received as a search result of all the web pages in which these search queries appear.

The co-occurrence rate calculating means 113 is a value obtained by subtracting the number of searches for “notebook personal computer AND optical drive” from the sum of the number of searches for “notebook personal computer” and “optical drive”. By dividing the number of searches, the co-occurrence rate of “notebook PC” and “optical drive” is calculated.
Further, the co-occurrence rate calculating means 113 searches for “notebook computer AND mouse” by subtracting the number of searches for “notebook computer AND mouse” from the sum of the number of searches for “notebook computer” and “mouse”. By dividing the number of cases, the co-occurrence rate of “notebook computer” and “mouse” is calculated.

  Returning to FIG. 1, the purchase candidate product group narrowing means 114 then selects a plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculation means 113 is a predetermined value or less (for example, 5% or less) from the purchase candidate product group. Exclude. That is, the plurality of subsequent purchase candidate product groups stored in the purchase log storage means 117 includes products purchased by the same user within a predetermined period, and is not suitable for recommending as a purchased product. It is considered that no combination is included. Then, the buy candidate product group narrowing means 114 eliminates those co-occurrence rates on the Web page that are less than or equal to the predetermined value from the buy candidate product groups, so that the co-occurrence rate exceeds the predetermined value. At the same time, a combination of products that are often posted on a web page can be accurately and then left as a candidate product group for purchase.

The recommendation list generation unit 115 stores the candidate purchase product group in the recommendation list storage unit 118 as a recommendation list after the purchase candidate product group narrowing unit 114 narrows down.
Further, the recommendation list generating unit 115 refers to the purchase log DB (see FIG. 2) stored in the purchase log storage unit 117, and then selects a product having the highest price among a plurality of products included in the candidate purchase product group. Set as the main product.
Further, the recommendation list generation unit 115 stores the weighting value (see FIG. 3) associated with the purchase candidate product group by the purchase candidate product group extraction unit 112 in the recommendation list and further associated with the purchase product group.

The recommendation list will be described with reference to FIG.
FIG. 5 is a diagram for explaining the recommendation list.
In the recommendation list shown in FIG. 5, the recommendation list generation unit 115 refers to the purchase log DB (see FIG. 2), and then selects “notebook computer” having the highest price among a plurality of products included in the candidate purchase product group. It is set as a product.
In the recommendation list, the main product ID is associated with the main product name, the purchased product ID, the purchased product name, and the unit price and score of the purchased product.
The main product ID is an identifier unique to each product for identifying the main product.
The main product name is the name of the main product type.
The purchased product ID is an identifier unique to each product for identifying the purchased product.
The purchased product name is the name of the type of the purchased product. In the recommendation list, each purchase product ID is associated with the purchase product name having the purchase product ID.
The unit price is the unit price of the purchased product. In the recommendation list, each purchased product ID is associated with the unit price of the purchased product having the next purchased product ID.
The score is a weighting value set by the buying candidate product group extraction unit 112. In the recommendation list, each purchased product ID is associated with the score of the next purchased product having the next purchased product ID. Also, in the recommendation list, for each main product, the purchased products are stored in descending order of score.
The score may be a value obtained by multiplying the weighting value set by the buying candidate commodity group extracting unit 112 and the co-occurrence rate calculated by the co-occurrence rate calculating unit 113.

Returning to FIG. 1, the purchased product recommendation device 12 includes a settlement request receiving unit 121, a reference amount arrival determining unit 122, a purchased product extracting unit 123, and a purchased product recommendation information transmitting unit 124.
The settlement request receiving unit 121 receives the settlement request transmitted from the user terminal 20. The payment request is information for selecting a product that the user desires to purchase on the Web page for selecting a product displayed on the user terminal 20 and requesting payment for the product. The payment request includes information such as the product ID, name, quantity, and unit price of the product for which payment is requested.

  The reference amount arrival determination unit 122 multiplies the quantity and the unit price of the product included in the payment request received by the payment request reception unit 121 to calculate the total price of the product included in the payment request. It is determined whether or not the reference amount has been reached. The predetermined reference amount is, for example, 100,000 yen when providing a discount service when the total amount of products is 100,000 yen or more.

  Next, when the purchase product extraction unit 123 determines that the predetermined reference amount has not been reached in the determination by the reference amount arrival determination unit 122, the purchase list (see FIG. 5) stored in the recommendation list storage unit 118 is used. Next, the purchased product is extracted. Specifically, the purchase product extraction unit 123 then refers to the recommendation list, searches for a main product that is the same type of product as the product included in the payment request, and is associated with the main product ID of the main product. Next, the product ID, name, unit price and score of the purchased product are extracted.

  Next, the purchased product recommendation information transmitting unit 124 transmits the product ID, name, and unit price of the purchased product extracted by the purchased product extracting unit 123 to the user terminal 20 that has transmitted the settlement request as the purchased product recommendation information. Specifically, the purchase merchandise recommendation information transmission means 124 then displays on the Web page requesting the settlement, for example, in addition to displaying that the total price of the merchandise is 100,000 yen or more and provides a discount. Next, recommendation information for displaying the product ID, name, and unit price is transmitted in descending order of the scores of the purchased products.

The user terminal 20 includes input means 210, display means 220, and user terminal control means 230.
The input unit 210 receives a click operation or the like in which a user selects a product and requests settlement on a Web page for selecting the product.
The display unit 220 displays a Web page for selecting a product and a Web page for requesting payment.
The user terminal control unit 230 controls the user terminal 20 and transmits information (for example, a payment request) accompanying the user operation received by the input unit 210 to the recommendation device 10. In addition, the user terminal control unit 230 receives information for displaying the product ID, name, and unit price of the purchased product in the Web page requested for settlement transmitted from the recommendation device 10 and causes the display unit 220 to display the information. Take control.

  The search server 30 includes a search engine and provides a search service. Specifically, the search result for the search query transmitted from the co-occurrence rate calculating unit 113 of the recommendation device 10 is transmitted to the co-occurrence rate calculating unit 113.

[Hardware configuration of recommendation system 1]
The recommendation device 10, the user terminal 20, and the search server 30 of the recommendation system 1 according to the embodiment are applied to a computer and its peripheral devices. Each unit in the recommendation device 10, the user terminal 20, and the search server 30 is configured by hardware included in the computer and its peripheral devices, and software that controls the hardware.

  The hardware includes a CPU (Central Processing Unit), a storage unit, a communication device, a display device, and an input device. Examples of the storage unit include a memory (RAM: Random Access Memory, ROM: Read Only Memory, etc.), a hard disk drive (HDD: Hard Disk Drive), and an optical disk (CD: Compact Disc, DVD: Digital Versatile Drive, etc.). Can be mentioned. Examples of the communication device include various wired and wireless interface devices. Examples of the display device include various displays such as a liquid crystal display and a plasma display. Examples of the input device include an input key, a touch panel, a keyboard, and a pointing device (mouse, tracking ball, etc.).

  The software includes a computer program and data for controlling the hardware. The computer program and data are stored in the storage unit, and are appropriately executed and referenced by each control unit. The computer program and data can be distributed via a communication line, or can be recorded on a computer-readable medium such as a CD-ROM and distributed.

[Control Flow of Recommendation Device 10]
Next, the control flow of the recommendation device 10 will be described.
FIG. 6 is a flowchart of processing in which the recommendation list generation device 11 of the recommendation device 10 according to the present embodiment generates a recommendation list.

  In step S11, a candidate product group extraction process is then performed. In this process, the purchase candidate commodity group extraction unit 112 (see FIG. 1) then refers to the purchase log DB (see FIG. 2) stored in the purchase log storage unit 117 (see FIG. 1) and associates a predetermined number of times or more. Then, a combination of a plurality of products is extracted as a purchase candidate product group. Further, in this process, the purchase candidate product group extraction unit 112 then refers to the purchase log stored in the purchase log storage unit 117 and sets a larger weight value as the number of times a plurality of products are associated with each other. Next, it is stored in association with the candidate product group for purchase.

  In step S12, a co-occurrence rate calculation process is performed. In this process, the co-occurrence rate calculation means 113 (see FIG. 1) uses the product names of a plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extraction means 112 in step S11 as a search query. A page is searched, and a co-occurrence rate that is a ratio of these search queries appearing on the same web page among all the web pages in which these search queries appear is calculated.

  In step S13, a candidate product group narrowing process is performed. In this process, the buying candidate product group narrowing means 114 (see FIG. 1) then excludes a plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculating unit 113 in step S12 is a predetermined value or less from the buying candidate product group. To do.

In step S14, a recommendation list generation process is performed. In this process, the recommendation list generating means 115 (see FIG. 1) is selected in the recommendation list storage means 118 (see FIG. 1) as a recommendation list after the purchase candidate commodity group narrowing means 114 narrows down in step S13. Remember.
In this process, the recommendation list generation unit 115 refers to the purchase log DB (see FIG. 2) stored in the purchase log storage unit 117, and then the price of a plurality of products included in the purchase candidate product group. The highest item is set as the main product.
Further, in this process, the recommendation list generation unit 115 uses the weighting value (see FIG. 3) associated with the purchase candidate product group in step S11 and the purchase candidate product group extraction unit 112 in the recommendation list. It is further associated with and stored.

  FIG. 7 is a flowchart of a process in which the purchased product recommendation device 12 subsequently transmits the recommendation information of the purchased product to the user terminal 20 after the recommendation device 10 according to the present embodiment.

  In step S21, payment request reception processing is performed. In this process, the settlement request receiving unit 121 (see FIG. 1) receives the settlement request transmitted from the user terminal 20 (see FIG. 1).

  In step S22, a reference amount arrival determination process is performed. In this process, the reference amount arrival determination means 122 (see FIG. 1) multiplies the quantity of the product included in the payment request received by the payment request reception means 121 in step S21 and the unit price, and calculates the product included in the payment request. The total amount is calculated, and it is determined whether or not the total amount has reached a predetermined reference amount. When it is determined that the total amount has reached the predetermined reference amount, the reference amount arrival determination means 122 ends this processing, and when it is determined that the total amount has not reached the predetermined reference amount The process moves to step S23.

  In step S23, a purchased product extraction process is performed. In this process, the purchased product extraction unit 123 (see FIG. 1) then refers to the recommendation list (see FIG. 5) stored in the recommendation list storage unit 118 (see FIG. 1), and then extracts the purchased product.

  In step S24, purchase product recommendation information transmission processing is then performed. In this process, the purchased product recommendation information transmitting means 124 (see FIG. 1) then displays the product ID, name, and unit price in descending order of the scores of the purchased products extracted by the purchased product extracting means 123 in step S23. For this reason, the recommended information is transmitted to the user terminal 20 that has transmitted the settlement request.

According to this embodiment, there are the following effects.
According to the present embodiment, the recommendation list generating device 11 generates a recommendation list for transmitting recommendation information of a purchased product that is purchased in combination with the main product to the user terminal 20. The purchase log storage unit 117 stores a plurality of products purchased by the same user within a predetermined period in the purchase log DB as purchase logs. Next, the purchase candidate product group extraction unit 112 refers to the purchase log DB stored in the purchase log storage unit 117 and then extracts a combination of a plurality of products associated with a predetermined number of times as a purchase candidate product group. Next, the co-occurrence rate calculating unit 113 searches the Web page using the product names of a plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extracting unit 112, and these search queries appear. A co-occurrence rate, which is a ratio of these search queries appearing on the same Web page among all Web pages, is calculated. Next, the purchase candidate product group narrowing-down unit 114 excludes a plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculating unit 113 is a predetermined value or less from the purchase candidate product group. The recommendation list generation unit 115 stores the candidate purchase product group in the recommendation list storage unit 118 as a recommendation list after the purchase candidate product group narrowing unit 114 narrows down.

As a result, products that are purchased at the same time as the main product, the number of purchases of which is equal to or greater than the predetermined number of times are extracted, and then the candidate product group is extracted from the group of candidate products for purchase. A recommendation list for transmitting recommendation information of the purchased product is generated after the product is purchased in combination with the main product by excluding the product whose value is equal to or less than the predetermined value. That is, the candidate product group to be purchased after the recommendation list is purchased at the same time as the main product and is posted on the same Web page more often than other products. Therefore, for example, when purchasing another product in addition to the main product, the user can select an appropriate purchase product without performing a complicated operation if a candidate purchase product group is displayed next. .
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

In addition, the recommendation list generation unit 115 refers to the purchase log DB stored in the purchase log storage unit 117, and then sets, as a main product, the one with the highest price among a plurality of products included in the purchase candidate product group. .
Thereby, since the main product has the highest unit price among a plurality of products purchased by the same user during the predetermined period, it is highly likely that the main product was the main purchase item of the user. And it is highly possible that the product other than the main product was an accessory or consumable of the main product. Thus, by determining the main product, it is possible to generate a recommendation list that matches the user's shopping mode.
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

Further, the recommendation list generation unit 115 stores the weighting value associated with the purchase candidate product group by the purchase candidate product group extraction unit 112 in the recommendation list and further associated with the purchase product group.
Thereby, since it is possible to rank the purchase candidate product group in the recommendation list by the weighting value, for example, it is possible to show the user in order from the purchase product having the higher weighting value.
Therefore, it is possible to provide a recommendation list generation device that creates a recommendation list for recommending a purchase product that is appropriate for the user to purchase in combination with the main product.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

11 Recommendation List Generation Device 117 Purchase Log Storage Unit 112 Next Purchase Candidate Product Group Extraction Unit 113 Co-occurrence Rate Calculation Unit 114 Next Purchase Candidate Product Group Narrowing Unit 115 Recommendation List Generation Unit

Claims (3)

A recommendation list generation device for generating a recommendation list for transmitting recommendation information of a purchased product to be purchased in combination with a main product to a user terminal connectable via a network,
Purchase log storage means for associating and storing a plurality of products purchased by the same user within a predetermined period as purchase logs;
Referring to the purchase log stored in the purchase log storage means, and then extracting a combination of the plurality of products associated with a predetermined number of times or more as a purchase candidate product group;
Next, a Web page is searched using the product names of the plurality of products included in the purchase candidate product group extracted by the purchase candidate product group extraction unit, and among all the Web pages in which these search queries appear. A co-occurrence rate calculating means for calculating a co-occurrence rate, which is the rate at which these search queries appear on the same Web page;
Next, the candidate product group narrowing means for excluding the plurality of products whose co-occurrence rate calculated by the co-occurrence rate calculating unit is a predetermined value or less from the buy candidate product group;
A recommendation list generation device comprising: a recommendation list generation unit that stores the candidate purchase product group as the recommendation list after the purchase candidate product group narrowing unit narrows down. The purchase log storage means further stores the price of the product in association with the product,
The said recommendation list production | generation means refers to the said purchase log memory | storage means, Then, the thing with the said highest price among the said some goods contained in the said purchase candidate goods group is set as main goods. Recommendation list generator. The purchase candidate product group extraction means then refers to the purchase log, sets a larger weighting value as the number of times the plurality of products are associated with each other, and stores the weighting value in association with the purchase candidate product group.
The recommendation list generation means stores the weight value associated with the purchase candidate commodity group by the purchase candidate commodity group extraction means and then further associates with the purchase commodity group in the recommendation list. The recommendation list generator described in 1.

Images