JP6509028B2 - Item recommendation device, item recommendation method and program - Google Patents

Description

  The present invention relates to a technology for recommending a product.

  Conventionally, collaborative filtering is known as one of product recommendation methods. Collaborative filtering is a method of recommending a product purchased by another user who has purchased those products to a user who is the product recommendation target, based on the products purchased by the product recommendation target user in the past. For example, in Patent Document 1, an adjacency matrix is generated from the customer's purchase history data, and by analyzing the adjacency matrix, the time-series order relationship between the products related to purchase is acquired, and the customer purchases the products. In doing so, it is described to recommend to the customer a product having a time-sequential order relationship with the product. In Patent Document 2, the order relationship of purchase between products and the average value of the time difference between purchases are obtained based on the purchase history of the customer, and when the customer purchases a certain product, the order relationship with the product is obtained. It is described that the direct mails of certain other products are to be shipped at the shipping timing determined based on the average value of the time difference of purchase of these products.

JP 2002-279279 A JP-A-9-305571

  A method using RWR (Random Walk with Restart) is known as one of product recommendation methods using collaborative filtering. RWR is a method of calculating the degree of association between nodes in graph theory, and is characterized in that the degree of association can be calculated based on the structural features of the graph. The product recommendation method using RWR has the property of being stronger in data sparsity than the method using machine learning, which is a typical example of collaborative filtering, and a method using singular value decomposition even in the method using the same matrix There is an advantage that calculation is lighter than it.

  However, the product recommendation method using RWR has a problem that it is not suitable for a product recommendation system with a large number of customers and recommendation-targeted product items, which is frequently updated. This is because it is necessary to re-do the calculation for constructing the graph structure of the product each time the purchase history data of the product used to generate the ranking of the product is updated. As the number of customers and the number of product items to be recommended increase, the product graph becomes larger, and proportionally, it takes time to update the ranking.

  The present invention has been made in view of such circumstances, and an object thereof is to reduce the amount of calculation required when recommending a product to a user.

  In order to solve the above problems, according to the present invention, there are a plurality of product recommendation degree storage units that store user identification information, product identification information, and recommendation levels in association with one user identification information. Product identification information, recommendation degree acquiring unit for acquiring a recommendation degree associated with each of the plurality of product identification information, user identification information, product identification information, and usage date and time are stored in association with each other Each of a first usage history acquisition unit that acquires usage dates and times that are associated with each of the plurality of item identification information items from the item usage history storage unit, and two item identification information items that constitute the plurality of item identification information items Regarding the set, the use interval from the first product identified by the first product identification information to the second product identified by the second product identification information is acquired after the first product identified by the first product identification information is used. Based on the date of use Among the plurality of product identification information items based on the usage interval identifying unit for identifying, the recommendation degree correcting unit for correcting the acquired recommendation degree based on the identified usage interval, and the corrected recommendation degree And a recommended product presentation unit that presents at least a portion of the above to the user identified by the one user identification information.

  In a preferable aspect, the recommendation degree correction unit is configured to set a recommendation degree corresponding to the second item identification information based on a use interval from when the first item is used to when the second item is used. To correct.

  In a further preferable aspect, the usage interval identification unit identifies a distribution of frequencies of usage intervals from the usage of the first product to the usage of the second product, and the recommendation degree correction unit The recommendation degree corresponding to the second item identification information is corrected based on the frequency of the use interval corresponding to the time interval elapsed from the last use of the first product to the present.

  In a further preferable aspect, the first usage history acquisition unit is a usage associated with any one of a plurality of pieces of user identification information including the one piece of user identification information and each of the plurality of item identification information The date and time is acquired from the product usage history storage unit.

  In a further preferable aspect, the information processing apparatus further comprises a second usage history acquisition unit that acquires, from the product usage history storage unit, usage dates and times associated with the one piece of user identification information and each of the plurality of item identification information; The use interval specifying unit is a use interval from when the first product is used to when the second product is used for each set of two item identification information that configures the plurality of item identification information. Is specified based on the use date and time acquired by the first use history acquisition unit, and the first goods are used for each set of two goods identification information constituting the plurality of goods identification information The usage interval from when the second product is used until the second product is used is specified based on the usage date and time acquired by the second usage history acquisition unit.

  Further, the present invention is a product recommendation method executed by a product recommendation device, wherein the product recommendation degree storage unit stores user identification information, product identification information, and a recommendation degree in association with each other. Obtaining a plurality of product identification information associated with the person identification information, a recommendation degree associated with each of the plurality of product identification information, user identification information, product identification information, and use date and time Each set consisting of a step of acquiring usage date and time correlated with each of the plurality of product identification information from the product usage history storage unit to be stored in association, and two pieces of product identification information constituting the plurality of product identification information The utilization interval from when the first item identified by the first item identification information is used to when the second item identified by the second item identification information is used is acquired. Advantage At least one of the plurality of product identification information items based on the date and time, the step of correcting the acquired recommendation degree based on the specified usage interval, and the corrected recommendation degree. And C. presenting a part to the user identified by the one user identification information.

  Further, according to the present invention, a plurality of product identifications associated with one user identification information from a product recommendation degree storage unit that stores user identification information, product identification information, and a recommendation degree in association with a computer. From the product usage history storage unit that stores information and a recommendation degree associated with each of the plurality of product identification information, user identification information, product identification information, and usage date and time in association with each other The step of acquiring a use date and time associated with each of the plurality of item identification information, and each pair of two item identification information constituting the plurality of item identification information is identified by the first item identification information; Identifying a use interval from the first product being used until the second product identified by the second product identification information is used based on the acquired use date and time Correcting the acquired recommendation degree based on the specified usage interval; and identifying at least a portion of the plurality of product identification information based on the corrected recommendation degree as the one user identification Providing a program for performing the steps of presenting to a user identified by the information.

  According to the present invention, it is possible to reduce the amount of calculation required when recommending a product to a user.

It is a figure which shows an example of a structure of the product recommendation system. FIG. 1 is a diagram showing an example of the configuration of a product recommendation device 1; It is a figure showing an example of purchase data DB12. It is a figure which shows an example of graph structure. It is a figure which shows an example of ranking DB14. It is a figure which shows an example of shop recommendation goods DB15. It is a figure which shows an example of the histogram which shows distribution of the frequency of the purchasing interval of goods of a customer individual. It is a figure which shows an example of the value of the probability calculated from the frequency of the purchasing interval of goods of a customer individual. It is a figure which shows an example of the graph of the value of the probability calculated from the frequency of the purchasing interval of a customer's individual goods. It is a figure which shows an example of the histogram which shows distribution of the frequency of the purchasing interval of goods of the whole customer. It is a figure which shows an example of the value of the probability calculated from the frequency of the purchasing interval of goods of the whole customer. It is a figure which shows an example of the graph of the value of the probability calculated from the frequency of the purchasing interval of goods of the whole customer. It is a figure which shows an example of addition of an individual space | interval score and the whole space | interval score. It is a figure which shows an example of a purchase interval score. It is a figure which shows an example of the graph of a purchase interval score. It is a flowchart which shows an example of a ranking production | generation process. It is a flow chart which shows an example of ranking amendment processing. It is a flow chart which shows an example of purchase interval reflection processing. It is a flow chart which shows an example of individual interval score calculation processing. It is a flow chart which shows an example of whole interval score calculation processing.

1. Embodiment 1-1. Configuration of Product Recommendation System 100 FIG. 1 is a view showing an example of the configuration of a product recommendation system 100 according to an embodiment of the present invention. As shown in FIG. 1, the product recommendation system 100 includes a product recommendation device 1, a POS (Point of Sale) terminal 2, and a communication terminal 3. These devices are connected to each other via the communication line 4. The communication line 4 is configured by, for example, the Internet, a mobile phone network, or a wireless LAN (Local Area Network). Although only one POS terminal 2 and one communication terminal 3 are shown in FIG. 1 for convenience of explanation, a plurality of each may be included in the product recommendation system 100.

  The product recommendation device 1 is a server device that collects POS purchase information from the POS terminal 2 and recommends a product to a customer who owns the communication terminal 3 based on this information.

  The POS terminal 2 is a register machine compatible with a POS system, which is installed in a store such as a supermarket or a convenience store. The POS terminal 2 transmits POS purchase information to the product recommendation device 1. The POS purchase information includes a customer ID, purchase date and time, and a product ID.

  The communication terminal 3 is a computer device owned by a customer. Specifically, they are portable electronic devices such as smartphones, mobile phones, and tablet terminals. Alternatively, it is a stationary electronic device such as a personal computer. The communication terminal 3 requests the product recommendation device 1 for the information on the recommended product, and displays the information on the recommended product received from the product recommendation device 1 on the screen.

1-2. Configuration of Product Recommendation Device 1 FIG. 2 is a view showing an example of the configuration of the product recommendation device 1. As illustrated in FIG. 2, the product recommendation device 1 generally includes a purchase data acquisition unit 11, a purchase data database (hereinafter, “database” is referred to as “DB”) 12, a ranking generation unit 13, and a ranking DB 14. And a shop recommendation product DB 15, a ranking correction unit 16, and a recommended product presentation unit 17. These functions are realized, for example, by execution of a program by an arithmetic device such as a CPU (Central Processing Unit).

  The purchase data acquisition unit 11 stores the POS purchase information transmitted from the POS terminal 2 in the purchase data DB 12. Specifically, the customer ID, the purchase date and time, and the product ID are stored in the purchase data DB 12 in association with each other.

  The purchase data DB 12 is a database for storing POS purchase information transmitted from the POS terminal 2. Purchase data DB12 is an example of the "commodity use history storage unit" according to the present invention. FIG. 3 is a diagram showing an example of the purchase data DB 12. Each record which comprises purchase data DB12 is constituted by each field of customer ID, purchase date and time, and goods ID. The customer ID is identification information that identifies the user of the communication terminal 3 and who uses the store where the POS terminal 2 is installed. The product ID is identification information for identifying a product, and is, for example, a Japanese Article Number (JAN) code.

  The ranking generation unit 13 generates ranking information of products based on POS purchase information stored in the purchase data DB 12 and stores the generated ranking information in the ranking DB 14. The ranking generation unit 13 generates product ranking information based on POS purchase information within a predetermined period (for example, within the past one year) in a predetermined cycle (for example, one month cycle) for each customer. As shown in FIG. 2, the ranking generation unit 13 includes a graph construction unit 131, a weight assignment unit 132, and a rank value calculation unit 133.

  The graph construction unit 131 constructs a graph structure of the product based on the POS purchase information stored in the purchase data DB 12. Specifically, the purchase date and time corresponding to the customer whose rank information is to be generated and the product ID are acquired from the purchase data DB 12, and the graph structure of the product is constructed based on these pieces of information. At this time, the purchase date and time within a predetermined period (for example, within the past one year) and the corresponding product ID are acquired. The graph structure to be constructed is composed of a node representing a product and an interactive edge linking two products purchased on the same day.

  FIG. 4 is a diagram showing an example of the graph structure. The graph structure shown in FIG. 4 indicates that "bacon" and "minced meat", "bacon" and "ginseng", "bacon" and "cabbage" were purchased on the same day. In addition, it shows that "Ginseng" and "Soy sauce", "Ginseng" and "Asparagus" were purchased on the same day.

  The weight assigning unit 132 assigns a weight to each edge of the graph structure constructed by the graph construction unit 131. As a weight, a value of transition probability is given. Specifically, a value obtained by dividing the number of purchases of two commodities tied by an edge on the same day by the sum of the number of purchases of transition destination commodities to be tied is given as a weight to the edge.

  In the graph structure shown in FIG. 4, the number of times "one (time)" the "bacon" and "minced meat" were purchased on the same day as the weight of the edge between "bacon" and "minced meat" is the same day as "bacon" The number of times you have purchased "ground meat" The number of times you have purchased "Ginseng" on the same day as "Bacon" The number of times you have purchased "Cabbage" on the same day as "Bon" The same day The value “1/3” obtained by dividing the sum “3 (times)” is given. In addition, as another example, as a weight of the edge between "Ginseng" and "Soy Sauce", "1 (time)" the number of times "Ginseng" and "Soy Sauce" were purchased on the same day The number of times you have purchased “1 (time)” and the number of times you purchased “asparagus” the same day as “ginseng” “1 (time)” Sum “2 (time)” divided by “2 (time)” It has been granted.

  The rank value calculation unit 133 calculates the rank value of each product in the graph structure in which the weight is given to the edge by the weight giving unit 132. At this time, the rank value is calculated using RWR. Here, RWR is an algorithm that calculates the degree of association (or similarity) of each node adjacent to the node that is the starting point on the graph structure. The degree of association corresponds to the rank value in the present embodiment. In RWR, a random walk is started from an origin node a, randomly moved to an adjacent node with a probability proportional to an edge weight, and returned to the origin node a with a certain probability each time a node is reached. The steady state probability obtained at each node as a result of iteratively repeating this operation is taken as the relevance. That is, the degree of association seen from the node a of the node b on the graph structure is calculated as the steady state probability of the node b obtained by random walking from the node a. For the method of calculating the degree of association using this RWR, for example, “M. Gori, 1 other person,“ Item Rank: A Random-Walk Based Scoring Algorithm for Recommender Engines ”, Proc. IJCAI '07, p. 2766-2771 "It is described in.

  In the graph structure shown in FIG. 4, the rank value (importance in FIG. 4) of the origin node “bacon” is “1.0 × 0.9 (= 0.9)”. Further, as another example, the rank value (importance in FIG. 4) of “minced meat” when the number of times of repetition is 3 times is “1.0 × 0.9 × (1/3) (= 0.3)” is there. The number of times of recursive repetition is not limited to three, and the steady state probability obtained as the value converges is taken as the rank value.

  When the rank value calculation unit 133 calculates the rank value of each product, the ranking generation unit 13 associates the rank value of a predetermined number of high-ranked products (for example, products with the rank value within the top 20) with the product ID. And store it in the ranking DB 14. At this time, the customer ID, the last purchase date of the product, and the ranking of the product are stored in the ranking DB 14 in association with the rank value and the product ID. In addition, when the latest purchased product is included in the upper predetermined number of products, the product may be excluded and the lower products may be advanced. This is because the rank value is calculated high for the latest purchased product.

  Next, the ranking DB 14 is a database that stores the ranking information generated by the ranking generation unit 13. The ranking DB 14 is an example of the “product recommendation degree storage unit” according to the present invention. FIG. 5 is a diagram showing an example of the ranking DB 14. Each record which comprises ranking DB14 is comprised by each field of customer ID, goods ID, last purchase date, rank value, and ranking. Here, the last purchase date is the date when the item identified by the corresponding item ID was recently purchased (in other words, the date when it was last purchased). The rank value is a value indicating the degree of recommendation of a product, and is calculated by the rank value calculation unit 133. The ranking is the order of goods determined based on the corresponding rank value. The larger the corresponding rank value, the smaller the value. The ranking DB 14 stores, for example, information on products ranked 20th or lower for each customer.

  The store recommendation product DB 15 is a database for storing information of products to be recommended to the customer at the store where the POS terminal 2 is installed. Specifically, it is a database that stores correction values used to correct the rank value of the store recommendation product. FIG. 6 is a diagram showing an example of the shop recommendation product DB 15. As shown in FIG. Each record which comprises store recommendation goods DB15 is constituted by each field of goods ID, recommendation start time, recommendation end time, and amendment value. Here, the recommendation start date and the recommendation end date are respectively the start date and the end date of a period in which the corresponding correction value is added to the rank value of the product identified by the corresponding product ID. The correction value is a value to be added to the rank value of the product identified by the corresponding product ID.

  The information stored in the store recommendation product DB 15 is updated, for example, by the manager of the store in which the POS terminal 2 is installed via the communication line 4 using a communication device (not shown). The manager of the store decides, for example, a recommended product or a non-recommended product according to the location, season, and stock of the store. A positive value is set as the correction value for the recommended product, and a negative value is set as the correction value for the non-recommended product.

  The ranking correction unit 16 corrects the ranking information stored in the ranking DB 14. Specifically, a correction is made to reflect the purchase interval of the product and the recommendation of the store on the rank value included in the ranking information. The ranking correction unit 16 corrects the ranking information at a higher frequency than the generation of the ranking by the ranking generation unit 13. For example, if ranking generation is performed on a monthly basis, correction of ranking information is performed on a daily basis. Alternatively, it may be executed each time the user of the communication terminal 3 requests information provision of a recommended product. As shown in FIG. 2, the ranking correction unit 16 is mainly composed of a ranking acquisition unit 161, a purchase interval reflection unit 162, and a shop recommendation reflection unit 163.

  The ranking acquisition unit 161 acquires, from the ranking DB 14, ranking information corresponding to a customer ID to be corrected for ranking information. The ranking acquisition unit 161 is an example of the “degree of recommendation acquisition unit” according to the present invention. The customer ID may be acquired from the communication terminal 3 together with a request for providing information on a recommended product, or may be stored in advance in the product recommendation device 1 as a target for periodically providing information on a recommended product. The ranking acquisition unit 161 may acquire all the ranking information corresponding to the customer ID from the ranking DB 14, or only the upper predetermined number of ranking information (for example, the ranking information in the top 20 ranks) should be corrected It may be acquired as

  The purchase interval reflection unit 162 corrects the rank value included in the ranking information acquired by the ranking acquisition unit 161 based on the purchase interval of the product. Specifically, the rank value is corrected on the basis of the purchase interval of the product of the individual customer who is the target of the correction of the ranking information and the purchase interval of the product of the entire customer registered in the purchase data DB 12. As shown in FIG. 2, the purchase interval reflection unit 162 is generally configured by an individual interval score calculation unit 1621, an entire interval score calculation unit 1622, an interval score calculation unit 1623, and a rank value correction unit 1624. .

  The individual interval score calculation unit 1621 specifies the distribution of the frequency of the purchase interval of the product of the customer individual who is the target of the correction of the ranking information, and calculates the value of the probability based on each frequency. Here, the product purchase interval is a time interval for purchasing two different types of products. For example, it is a time interval after purchasing the product A until purchasing the product B at a later date. As described later, the purchase interval of the product may be a time interval for purchasing two products identified by the same product ID. The individual interval score calculation unit 1621 includes a histogram generation unit 16211 and a probability calculation unit 16212.

  The histogram creation unit 16211 creates a histogram indicating the distribution of the frequency of the purchase intervals of the product of the customer individual who is the target of the correction of the ranking information. The histogram creation unit 16211 is an example of the “second usage history acquisition unit” and the “use interval specification unit” according to the present invention. Specifically, the histogram creation unit 16211 first acquires the purchase date and time corresponding to the customer ID and each product ID included in the ranking information acquired by the ranking acquisition unit 161 from the purchase data DB 12. At this time, the purchase date and time within a predetermined period (for example, within the past one year) is acquired. Further, the histogram creation unit 16211 acquires from the purchase data DB 12 the latest purchase date and time corresponding to the customer ID to be corrected for the ranking information and the product ID corresponding to the purchase date and time. Here, the latest purchase date and time may be the latest purchase date and time, or may be one or more purchase date and time belonging to a predetermined period from the time of the histogram creation processing. Then, for the arbitrary combination of the product ID included in the ranking information acquired by the ranking acquisition unit 161 and the product ID corresponding to the latest purchase date and time, until the second product is purchased after the purchase of the first product. Create a histogram that shows the distribution of the frequency of the purchase intervals. For example, the first product is purchased as "2015/4/1" and "2015/4/8", and the second product is purchased as "2015/4/3" and "2015/4/14" If the second product is purchased after “2015/4/1” after the first product is purchased for “2015/4/1,” the purchase interval is “2 days”. The frequency increases by "1". In addition, after the first product is purchased for "2015/4/8", the second product is purchased for "2015/4/14", so the frequency of the purchase interval "6 days" is " 1 "increases. If the ranking information includes, for example, 20 product IDs, and the number of product IDs corresponding to the latest purchase date and time is 10, the histogram creation unit 16211 sets 200 (= 20 × 10). Create a histogram for each customer ID.

  FIG. 7 is a diagram showing an example of the histogram created by the histogram creation unit 16211. The horizontal axis of the histogram shown in FIG. 7 indicates the purchase interval (in weeks), and the vertical axis indicates the number of purchases. The histogram shows the distribution of the frequency of the purchase interval until the product B is purchased after the product A is purchased. This histogram indicates that, for example, after the purchase of the product A, the product B is purchased once in the second week (between the eighth day and the 14th day). Also, as another example, it indicates that the product B was purchased twice in the fifth week (between the 29th day and the 35th day) after the purchase of the product A.

  The probability calculation unit 16212 calculates the value of probability (personal interval score) based on each frequency of the histogram created by the histogram creation unit 16211. Specifically, each frequency is input to the kernel density function to calculate the value of probability. The kernel density function p (x) is expressed, for example, as follows.

ここで、Ｎは、処理対象の購入間隔の数であり、ｘ_iは、各購入間隔の度数である。バンド幅ｄは、例えば以下のように表される。

Here, N is the number of purchase intervals to be processed, and x _i is the frequency of each purchase interval. The bandwidth d is expressed, for example, as follows.

ここで、σは、処理対象の購入間隔の度数についての標準偏差である。

Here, σ is a standard deviation of the frequency of the purchase interval to be processed.

  FIG. 8 is a diagram illustrating an example of the value of the probability calculated by the probability calculating unit 16212. FIG. 9 is a graph of probability values. The horizontal axis of the graph shown in FIG. 9 indicates the purchase interval (in units of weeks), the vertical axis on the left indicates the number of purchases, and the vertical axis on the right indicates the probability. The probability calculation unit 16212 calculates the value of probability for the frequency of each histogram created by the histogram creation unit 16211.

  Next, the entire interval score calculation unit 1622 specifies the distribution of the purchase intervals of the products of the entire customer registered in the purchase data DB 12, and calculates the value of the probability based on each frequency. Here, the product purchase interval is a time interval for purchasing two different types of products. For example, it is a time interval after purchasing the product A until purchasing the product B at a later date. As described later, the purchase interval of the product may be a time interval for purchasing two products identified by the same product ID. The overall interval score calculation unit 1622 is different from the individual interval score calculation unit 1621 in that the probability calculation is performed based on the frequency of the purchase intervals of the products for the entire customer, not for a specific customer individual. The overall interval score calculation unit 1622 is configured of a histogram creation unit 16221 and a probability calculation unit 16222.

  The histogram creation unit 16221 creates a histogram indicating the distribution of the frequency of the purchase intervals of the products of the entire customer registered in the purchase data DB 12. The histogram creation unit 16221 is an example of the “first usage history acquisition unit” and the “use interval identification unit” according to the present invention. Specifically, the histogram creation unit 16221 first acquires, from the purchase data DB 12, the purchase date and time corresponding to each product ID included in the ranking information acquired by the ranking acquisition unit 161. At this time, the purchase date and time within a predetermined period (for example, within the past one year) is acquired. Further, the histogram creation unit 16221 acquires from the purchase data DB 12 the latest purchase date and time corresponding to all the customer IDs and the product ID corresponding to the purchase date and time. Here, the latest purchase date and time may be the latest purchase date and time, or may be one or more purchase date and time belonging to a predetermined period from the time of the histogram creation processing. Then, the histogram creating unit 16221 is configured to purchase the second product after purchasing the first product for any combination of the product ID included in the ranking information acquired by the ranking acquiring unit 161 and the product ID corresponding to the latest purchase date and time. Create a histogram that shows the distribution of the frequency of the purchase interval until you purchase the item. For example, the first product is purchased as "2015/4/1" and "2015/4/8", and the second product is purchased as "2015/4/3" and "2015/4/14" If the second product is purchased after “2015/4/1” after the first product is purchased for “2015/4/1,” the purchase interval is “2 days”. The frequency increases by "1". In addition, after the first product is purchased for "2015/4/8", the second product is purchased for "2015/4/14", so the frequency of the purchase interval "6 days" is " 1 "increases. If the ranking information includes, for example, 50 product IDs, and the number of product IDs corresponding to the latest purchase date is 50, the histogram creation unit 16221 sets 2500 (= 50 × 50). Create a histogram. If the number of product IDs corresponding to the latest purchase date is huge and the number of histograms to be created is large and it is difficult to calculate, the product ID to be used is the product ID of the top 20 most frequently purchased products. It may be limited to

  FIG. 10 is a diagram illustrating an example of the histogram created by the histogram creation unit 16221. The horizontal axis of the histogram shown in FIG. 10 indicates the purchase interval (in weeks), and the vertical axis indicates the number of purchases (more specifically, the cumulative value of the number of purchases for each customer). The histogram shows the distribution of the frequency of the purchase interval until the product B is purchased after the product A is purchased. This histogram indicates that, for example, after the purchase of the product A, the product B is purchased 29 times in the first week (between the first day and the seventh day). Also, as another example, it indicates that the product B was purchased 14 times in the second week (between the eighth day and the 14th day) after the purchase of the product A.

  The probability calculation unit 16222 calculates the value of the probability (overall interval score) based on each frequency of the histogram created by the histogram creation unit 16211. Specifically, each frequency is input to the above-mentioned kernel density function to calculate the value of probability.

  FIG. 11 is a diagram illustrating an example of the value of the probability calculated by the probability calculating unit 16222. FIG. 12 is a graph of probability values. The horizontal axis of the graph shown in FIG. 12 indicates the purchase interval (in weeks), and the vertical axis on the left indicates the number of purchases (more specifically, the cumulative value of the number of purchases for each customer). The axis shows the probability. The probability calculation unit 16222 calculates the value of probability for the frequency of each histogram created by the histogram creation unit 16221.

  Next, the interval score calculation unit 1623 calculates the probability value calculated based on each frequency of the histogram by the individual interval score calculation unit 1621 and the probability calculated based on each frequency of the histogram by the overall interval score calculation unit 1622 The purchase interval score is calculated by adding the value of and dividing the obtained value by "2". At this time, the interval score calculation unit 1623 adds two probability values for each purchase interval. Also, the interval score calculation unit 1623 performs this addition and division processing between corresponding probability value pairs. Specifically, it is a combination of the same products, which is performed between the set of probability values calculated for the combinations having the same purchase order. For example, after the product A is purchased, the product B is purchased after the combination of the value of the probability calculated by the individual interval score calculation unit 1621 based on each frequency of the purchase interval until the product B is purchased after the product A is purchased It carries out between sets of the value of the probability calculated by the whole space | interval score calculation part 1622 based on each frequency of the purchasing space until it carries out.

  FIG. 13 is a diagram illustrating an example of the addition performed by the interval score calculation unit 1623. In the example of addition shown in FIG. 13, the set of probability values in the upper row is a set of probability values calculated by the individual interval score calculation unit 1621 (see FIG. 8). That is, it is a set of probability values calculated based on the frequency of the purchase interval of the product of the individual customer who is the target of correction of the ranking information. The set of probability values is a set of probability values calculated on the basis of the frequency of the purchase interval from the purchase of the product A to the purchase of the product B. The middle set of probability values is a set of probability values calculated by the overall interval score calculation unit 1622 (see FIG. 11). That is, it is a set of probability values calculated based on the frequency of purchase intervals of products for the entire customer. The set of probability values is also a set of probability values calculated based on the frequency of the purchase interval from purchase of the product A to purchase of the product B. The lower set of probability values is a set of values calculated by adding the upper probability value and the lower probability value for each purchase interval.

  In the example of the addition shown in FIG. 13, for example, the value “0.156632” is calculated by adding the value “0.04182” of the probability of the upper stage of the first week and the value “0.10846” of the probability of the middle stage. Is shown. Further, as another example, it is shown that the value “0.212589” is calculated by adding the value “0.096845” of the probability in the upper row of the second week and the value “0.115745” of the probability in the middle.

  FIG. 14 is a diagram showing an example of the purchase interval score finally obtained by the interval score calculation unit 1623. FIG. 15 is a graph showing a purchase interval score. The horizontal axis of the graph shown in FIG. 15 indicates the purchase interval (in weeks), and the vertical axis indicates the probability. Each purchase interval score shown in FIG. 14 is a value obtained by dividing the value of the probability shown in the lower part of FIG. 13 by “2”. According to the graph shown in FIG. 15, the purchase interval score peaks in the fifth and ninth weeks, similarly to the probability value calculated by the individual interval score calculation unit 1621. Further, the purchase interval score is a value of the probability calculated by the individual interval score calculation unit 1621 under the influence of the value of the probability calculated by the entire interval score calculation unit 1622 in the first week and the second week. Its value is larger than that. That is, the purchase interval score reflects both the purchase interval of the customer's individual item and the purchase interval of the item for the entire customer.

  The rank value correction unit 1624 corrects the rank value included in the ranking information acquired by the ranking acquisition unit 161 based on the purchase interval score calculated by the interval score calculation unit 1623. The rank value correction unit 1624 is an example of the “degree of recommendation correction unit” according to the present invention. The rank value correction unit 1624 performs the correction based on each set of purchase interval scores calculated by the interval score calculation unit 1623.

  Specifically, first, the rank value correction unit 1624 determines the time interval between the last purchase of the preceding purchased product in the purchase interval of the product referred to for calculating the purchase interval score and the time of the correction process. Identify For example, when the rank value is corrected based on the purchase interval score calculated based on the frequency of the purchase interval after the purchase of the product A until the purchase of the product B, the date and time when the product A was recently purchased And a time interval from the date and time of the last purchase to the time of the correction process. Here, the date and time at which the product A was recently purchased is specified, for example, by referring to the previous purchase date and time associated with the product A in the ranking DB 14.

  Next, the rank value correction unit 1624 specifies a purchase interval score corresponding to the specified time interval. For example, if the specified time interval is “3 days”, and the purchase interval score to be referred to is as shown in the example shown in FIG. 14, the purchase interval score “0.078316” for the first week is specified. As another example, if the specified time interval is “9 days”, the purchase interval score “0.106295” for the second week shown in FIG. 14 is specified.

  Finally, the rank value correction unit 1624 multiplies the rank value of the subsequent purchased product in the purchase interval of the product referred to in order to calculate the purchase interval score by the specified purchase interval score. For example, when the rank value is to be corrected based on the purchase interval score calculated based on the frequency of the purchase interval until the purchase of the product B after the purchase of the product A, the rank value of the product B is specified Multiply purchase interval score.

  Next, the store recommendation reflecting unit 163 corrects each rank value corrected by the purchase interval reflecting unit 162 based on the recommendation of the store. Specifically, each rank value is corrected based on the information stored in the store recommendation product DB 15. For example, in the store recommendation product DB 15 shown in FIG. 6, since the correction value “+0.1” is set for the product ID “11111111”, the date and time of execution of the correction process is the recommendation start date “2015/5/1”. If it is included in the range of the end date of recommendation “2015/5/31”, “0.1” is added to the rank value corresponding to the product ID “11111111”.

  Next, the recommended product presentation unit 17 determines the ranking of each product based on the rank value corrected by the ranking correction unit 16, and presents the determined ranking together with the product ID to the user of the communication terminal 3. Under the present circumstances, the recommendation product presentation part 17 gives a ranking of a small value with respect to goods, so that the corresponding rank value is large. Further, specifically, the recommended product presentation unit 17 transmits, to the communication terminal 3, screen data for displaying the product ID and the ranking.

  The recommended product presentation unit 17 presents the product ID and the ranking to the user of the communication terminal 3 only for a predetermined number of high-ranked products (for example, products within the top 5 rankings) based on the determined ranking. You may do so. Further, the recommended product presentation unit 17 may present a product name, an image of a product, or a price of a product instead of or in addition to the product ID. In addition, when the product recommendation device 1 includes the display device, the recommended product presentation unit 17 displays the screen data on the display device without transmitting the screen data to the communication terminal 3, thereby the communication terminal 3. The product ID and the ranking may be presented to the user of.

1-3. Operation The operation of the product recommendation device 1 will be described. Specifically, ranking generation processing and ranking correction processing will be described.

1-3-1. Ranking Generation Process FIG. 16 is a flowchart showing an example of the ranking generation process. The following description assumes that ranking information for one customer is generated. The ranking generation process is performed by the ranking generation unit 13, for example, in a predetermined cycle (for example, a one-month cycle).

  In step Sa1 of the present process, the graph construction unit 131 constructs a graph structure of the product based on the POS purchase information stored in the purchase data DB 12. Specifically, the purchase date and time corresponding to the customer whose rank information is to be generated and the product ID are acquired from the purchase data DB 12, and the graph structure of the product is constructed based on these pieces of information. The graph structure to be constructed is composed of a node representing a product and an interactive edge linking two products purchased on the same day.

  Next, the weight assignment unit 132 assigns a weight to each edge of the graph structure constructed in step Sa1 (step Sa2). Specifically, a value obtained by dividing the number of purchases of two products tied by an edge by the number of purchases of a transition destination product is given as a weight to the edge.

  Next, the rank value calculation unit 133 calculates the rank value of each product in the graph structure in which the edge is given a weight in step Sa2 (step Sa3). At this time, the rank value is calculated using the above-mentioned RWR.

Next, the ranking generation unit 13 stores the rank value of each product calculated in step Sa3 in association with the product ID in the ranking DB 14 (step Sa4). At this time, only the rank value of a predetermined number of high-order products (for example, products having a rank value of the top 20 or less) may be stored in the ranking DB 14 in association with the product ID. When storing the rank value and the product ID, the customer ID, the last purchase date and time of the product, and the ranking of the product are stored in the ranking DB 14 in association with these.
The above is the description of the ranking generation process.

1-3-2. Ranking Correction Process FIG. 17 is a flowchart showing an example of the ranking correction process. The ranking correction process is executed by the ranking correction unit 16 more frequently than the ranking generation process. The ranking correction process is executed, for example, when a user of the communication terminal 3 requests to provide information on a recommended product.

  In step Sb1 of the present process, the ranking acquisition unit 161 acquires ranking information corresponding to a customer ID that is a target of correction of the ranking information from the ranking DB. The customer ID is acquired from the communication terminal 3 together with, for example, a request for providing information on a recommended product. Here, the ranking acquisition unit 161 may acquire only a predetermined number of upper ranking pieces of ranking information (for example, ranking information in the top twenty ranks) from the ranking DB 14 as a correction target.

  Next, the purchase interval reflection unit 162 corrects the rank value included in the ranking information acquired in step Sb1 based on the purchase interval of the product (step Sb2). Specifically, the rank value is corrected on the basis of the purchase interval of the product of the individual customer who is the target of the correction of the ranking information and the purchase interval of the product of the entire customer registered in the purchase data DB 12.

  FIG. 18 is a flowchart showing an example of the purchase interval reflection process performed in step Sb2. In step Sc1 of the present process, the individual interval score calculation unit 1621 specifies the distribution of the frequency of the purchase interval of the product of the customer individual who is the target of the correction of the ranking information, and calculates the probability value based on each frequency. Here, the product purchase interval is a time interval for purchasing two different types of products. As described later, the purchase interval of the product may be a time interval for purchasing two products identified by the same product ID.

  FIG. 19 is a flowchart showing an example of the individual interval score calculation process executed in step Sc1. In step Sd1 of the present process, the histogram creating unit 16211 creates a histogram indicating the distribution of the frequency of the purchase intervals of the products of the individual customer who is the target of the correction of the ranking information. Specifically, first, the purchase date and time corresponding to the customer ID and each product ID included in the ranking information acquired in step Sb1 are acquired from the purchase data DB 12. Further, the latest purchase date and time corresponding to the customer ID to be corrected for the ranking information and the product ID corresponding to the purchase date and time are acquired from the purchase data DB 12. Then, for any combination of the product ID included in the above ranking information and the product ID corresponding to the latest purchase date, the frequency of the purchase interval before the purchase of the second product after the purchase of the first product. Create a histogram showing the distribution.

  Next, the probability calculating unit 16212 calculates the value of probability based on each frequency of the histogram created in step Sd1 (step Sd2). Specifically, each frequency is input to the kernel density function to calculate the value of probability. The probability calculation unit 16212 calculates the value of probability for the frequency of each histogram created in step Sd1.

  In step Sc2 of FIG. 18, the entire interval score calculation unit 1622 specifies the distribution of the purchase intervals of the products of the entire customer registered in the purchase data DB 12, and calculates the value of the probability based on each frequency. Here, the product purchase interval is a time interval for purchasing two different types of products. As described later, the purchase interval of the product may be a time interval for purchasing two products identified by the same product ID.

  FIG. 20 is a flowchart showing an example of the overall interval score calculation process performed in step Sc2. In step Se1 of the present process, the histogram creation unit 16221 creates a histogram indicating the distribution of the frequency of the purchase intervals of the products of all the customers registered in the purchase data DB 12. Specifically, first, the purchase date and time corresponding to each product ID included in the ranking information acquired in step Sb1 is acquired from the purchase data DB 12. Further, the latest purchase date and time corresponding to all the customer IDs and the product ID corresponding to the purchase date and time are acquired from the purchase data DB 12. Then, for any combination of the product ID included in the above ranking information and the product ID corresponding to the latest purchase date, the frequency of the purchase interval before the purchase of the second product after the purchase of the first product. Create a histogram showing the distribution.

  The probability calculating unit 16222 calculates the value of probability based on each frequency of the histogram created in step Se1 (step Se2). Specifically, each frequency is input to the above-mentioned kernel density function to calculate the value of probability. The probability calculation unit 16222 calculates the value of probability for the frequency of each histogram created in step Se1.

  In step Sc3 of FIG. 18, the interval score calculation unit 1623 adds the value of the probability calculated in step Sc1 and the value of the probability calculated in step Sc2 and divides the obtained value by “2”. By doing this, the purchase interval score is calculated. At this time, the interval score calculation unit 1623 adds two probability values for each purchase interval. Also, the interval score calculation unit 1623 performs this addition and division processing between corresponding probability value pairs.

  Next, the rank value correction unit 1624 corrects the rank value included in the ranking information acquired in step Sb1 based on the purchase interval score calculated in step Sc3 (step Sc4). The rank value correction unit 1624 performs the correction based on each set of purchase interval scores calculated in step Sc3.

  Specifically, the rank value correction unit 1624 first specifies a time interval from the recent purchase of the previously purchased product in the purchase interval of the product referred to for calculating the purchase interval score to the time of the correction process. Do. Next, the rank value correction unit 1624 specifies a purchase interval score corresponding to the specified time interval. Finally, the rank value correction unit 1624 multiplies the rank value of the subsequent purchased product in the purchase interval of the product referred to in order to calculate the purchase interval score by the specified purchase interval score.

  Next, in step Sb3 of FIG. 17, the shop recommendation reflecting unit 163 corrects each rank value corrected in step Sb2 based on the recommendation of the shop. Specifically, each rank value is corrected based on the information stored in the store recommendation product DB 15.

Next, the recommended product presentation unit 17 determines the ranking of each product based on the rank value corrected in step Sb3, and presents the determined ranking together with the product ID to the user of the communication terminal 3 (step Sb4). Specifically, screen data for displaying the product ID and the ranking is transmitted to the communication terminal 3.
The above is the description of the ranking correction process.

  According to the product recommendation device 1 according to the present embodiment described above, when information on a recommended product is provided to the customer, ranking generation that requires recalculation using RWR is not performed, and only the ranking correction is performed. Is performed, thereby reducing the amount of calculation required to recommend a product to the customer. For this reason, the product recommendation device 1 according to the present embodiment can cope with a product recommendation system that has a large number of customers and product items to be recommended, and has a high update frequency.

  Further, according to the product recommendation device 1 according to the present embodiment, the ranking is corrected based on the purchase intervals of the products of each customer, so that the product may be recommended in accordance with the timing at which the customer is likely to purchase. it can. In addition, since the ranking is corrected based on the purchase intervals of the products for the entire customer as well as the purchase intervals of the products for each customer, even if there is a shortage of data on the purchase intervals of the products for the individual customers, the shortage is It can be supplemented by data of purchase intervals of products for the entire customer.

  Further, according to the product recommendation device 1 according to the present embodiment, since the ranking is corrected based on the recommendation of the store, the product can be recommended to the customer according to the location, the season, and the stock of the store.

2. Modifications The above embodiment may be modified as shown below. Further, the following modifications may be combined with each other.

2-1. Modification 1
In the above description of the embodiment, the food is taken as an example of the product (see, for example, FIG. 4), but this is only an example of the product. As used herein, a product is a good or service that is the object of a commercial transaction. Here, the goods may be real or intangible (for example, digital content). In addition, commercial transactions include not only buying and selling but also rental (rental). That is, the product recommended to the customer in the above embodiment may be an object of rent.

2-2. Modification 2
Although the product recommendation method using RWR is used in the ranking generation processing according to the above embodiment, this is only an example of a ranking generation method. In the above ranking generation process, other conventional collaborative filtering (eg, “Sarwar, 3 others,“ Item-based Collaborative Filtering Recommendation Algorithms ”, Proc. WWW '01, p.285-295 ”may be used to generate ranking information.

2-3. Modification 3
In the ranking correction process according to the above embodiment, the kernel density function is used to calculate the value of probability based on the frequency of the histogram, but this is only one example of means for calculating the value of probability. . In the above ranking correction processing, the interval score may be calculated by applying the frequency of the histogram to an exponential distribution or a normal distribution, instead of the kernel density function.

2-4. Modification 4
In the ranking correction process according to the above embodiment, when correcting the rank value based on the purchase interval of the products, reference is made to the time interval for purchasing two different types of products, but instead, the same type The time interval for purchasing the product may be referred to. Specifically, the rank value may be corrected based on the time interval from the purchase of the product A to the purchase of the product A next time.

  In this case, in the above embodiment, the histogram creating unit 16211 and the histogram creating unit 16221 purchase the product for each product ID included in the ranking information acquired by the ranking acquisition unit 161 and then the product Create a histogram that shows the distribution of the frequency of the purchase interval until purchase. The interval score calculation unit 1623 is a value obtained by adding the value of the probability calculated by the individual interval score calculation unit 1621 and the value of the probability calculated by the overall interval score calculation unit 1622 for the same kind of product. The purchase interval score is calculated by dividing by “2”. The rank value correction unit 1624 specifies a time interval from the last purchase of the product corresponding to the purchase interval score to the time of the correction process, specifies the purchase interval score corresponding to the specified time interval, and The product rank value is multiplied by the specified purchase interval score.

  Also according to this modification, it is possible to recommend a product in accordance with the timing at which the customer is likely to purchase.

2-5. Modification 5
In the ranking correction process according to the above-described embodiment, the process of correcting the rank value based on the recommendation of the store by the store recommendation reflection unit 163 may be omitted. Further, in the ranking correction processing according to the above embodiment, the purchase interval reflection unit 162 corrects the rank value based on only one of the purchase interval of the product of the customer individual and the purchase interval of the product of the entire customer. You may do it. That is, rank value correction unit 1624 sets the rank value based on only one of the probability value calculated by individual interval score calculation unit 1621 and the value of the probability calculated by whole interval score calculation unit 1622. May be corrected.

2-6. Modification 6
In the above embodiment, the product recommendation device 1 includes the purchase data acquisition unit 11, the purchase data DB 12, the ranking generation unit 13, the ranking DB 14, and the store recommendation product DB 15, but these constituent elements are It may be provided in another device connected to the product recommendation device 1 via the communication line 4.

2-7. Modification 7
The program executed in the product recommendation device 1 according to the above embodiment may be provided via a recording medium readable by a computer device. Here, the recording medium is, for example, a magnetic recording medium such as a magnetic tape or a magnetic disk, an optical recording medium such as an optical disk, a magneto-optical recording medium, a semiconductor memory or the like. Also, the program may be distributed via a network such as the Internet.

DESCRIPTION OF SYMBOLS 1 ... commodity recommendation device, 2 ... POS terminal, 3 ... communication terminal, 4 ... communication line, 11 ... purchase data acquisition part, 12 ... purchase data DB, 13 ... ranking generation part, 14 ... ranking DB, 15 ... store recommendation product DB, 16: ranking correction unit, 17: recommended product presentation unit, 100: product recommendation system, 131: graph construction unit, 132: weighting unit, 133: rank value calculation unit, 161: ranking acquisition unit, 162: purchase interval Reflecting part, 163: store recommendation reflecting part, 1621: personal spacing score calculating part, 1622: whole spacing score calculating part, 1623: spacing score calculating part, 1624: rank value correcting part, 16211 ... histogram creating part, 16212: probability calculating Part, 16221 ... Histogram creation part, 16222 ... Probability calculation part

Claims (7)

A plurality of item identification information items associated with one user identification information item from the item recommendation degree storage unit that stores user identification information items, item identification information items, and recommendation degrees in association with one another, and the plurality of item item identification information items A recommendation degree acquisition unit that acquires a recommendation degree associated with each of the
The first usage history acquisition for acquiring the usage date and time associated with each of the plurality of product identification information from the product usage history storage unit that stores the user identification information, the product identification information, and the usage date in association with each other Department,
The first product identified by the first product identification information is used for each set of the two product identification information that make up the plurality of product identification information, and is then identified by the second product identification information A usage interval specifying unit that specifies, based on the acquired usage date and time, a usage interval until the second product is used;
A recommendation degree correction unit that corrects the acquired recommendation degree based on the specified usage interval;
A recommended product presentation unit which presents at least a part of the plurality of product identification information to the user identified by the one user identification information based on the corrected degree of recommendation .   The recommendation degree correction unit corrects the recommendation degree corresponding to the second item identification information based on a use interval from use of the first item to use of the second item. The product recommendation device according to claim 1, characterized in that The use interval specifying unit specifies a distribution of frequencies of use intervals from the use of the first product to the use of the second product,
The recommendation degree correction unit is configured to set the recommendation degree corresponding to the second item identification information based on the frequency of the use interval corresponding to the time interval elapsed from the last use of the first product to the present The product recommendation device according to claim 2, wherein the correction is made.   The first usage history acquisition unit uses the product date and time of use, which is associated with any one of a plurality of pieces of user identification information including the one piece of user identification information and each of the plurality of product identification information. The product recommendation device according to any one of claims 1 to 3, wherein the product recommendation device is acquired from a history storage unit. It further comprises a second usage history acquisition unit for acquiring from the product usage history storage unit the usage date and time associated with the one piece of user identification information and each of the plurality of product identification information,
The use interval specifying unit is a use interval from when the first product is used to when the second product is used for each set of two item identification information that configures the plurality of item identification information. Is specified based on the use date and time acquired by the first use history acquisition unit, and the first goods are used for each set of two goods identification information constituting the plurality of goods identification information The product recommendation according to claim 4, wherein a use interval from when the second product is used until the second product is used is specified based on the use date and time acquired by the second use history acquisition unit. apparatus. A product recommendation method executed by a product recommendation device, comprising:
A plurality of item identification information items associated with one user identification information item from the item recommendation degree storage unit that stores user identification information items, item identification information items, and recommendation degrees in association with one another, and the plurality of item item identification information items Obtaining a degree of recommendation associated with each of
Acquiring a use date and time associated with each of the plurality of product identification information from a product use history storage unit that stores user identification information, product identification information, and use date in association with each other;
The first product identified by the first product identification information is used for each set of the two product identification information that make up the plurality of product identification information, and is then identified by the second product identification information Specifying a use interval until the second product is used based on the acquired use date and time;
Correcting the acquired recommendation degree based on the specified usage interval;
Presenting at least a part of the plurality of product identification information to the user identified by the one user identification information based on the corrected recommendation degree. On the computer
A plurality of item identification information items associated with one user identification information item from the item recommendation degree storage unit that stores user identification information items, item identification information items, and recommendation degrees in association with one another, and the plurality of item item identification information items Obtaining a degree of recommendation associated with each of
Acquiring a use date and time associated with each of the plurality of product identification information from a product use history storage unit that stores user identification information, product identification information, and use date in association with each other;
The first product identified by the first product identification information is used for each set of the two product identification information that make up the plurality of product identification information, and is then identified by the second product identification information Specifying a use interval until the second product is used based on the acquired use date and time;
Correcting the acquired recommendation degree based on the specified usage interval;
A program for causing at least a part of the plurality of product identification information to be presented to a user identified by the one user identification information based on the corrected recommendation degree.

Images