WO2016092767A1 - グルーピングシステム、グルーピング方法およびグルーピングプログラム - Google Patents
グルーピングシステム、グルーピング方法およびグルーピングプログラム Download PDFInfo
- Publication number
- WO2016092767A1 WO2016092767A1 PCT/JP2015/005926 JP2015005926W WO2016092767A1 WO 2016092767 A1 WO2016092767 A1 WO 2016092767A1 JP 2015005926 W JP2015005926 W JP 2015005926W WO 2016092767 A1 WO2016092767 A1 WO 2016092767A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- product
- customer
- group
- purchase
- grouping
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0204—Market segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Definitions
- the present invention relates to a grouping system, a grouping method, and a grouping program for grouping customers and grouping items or items related to items or items related to services or services.
- Non-Patent Document 1 describes a method for creating customer clusters.
- characteristics such as “low price” and “high-grade” are given to each product in advance.
- the feature imparted to the product may be referred to as product DNA.
- the characteristic provided to the goods which each customer purchased is totaled, and the result is made into the characteristic of a customer. This feature is sometimes referred to as customer DNA.
- a cluster of customers is created based on the customer characteristics.
- Non-Patent Document 2 describes a method for classifying customers.
- product DNA such as “easy health system” and “time saving system” representing a lifestyle is determined and applied to various products. For example, a customer who purchases a certain amount of a product called “Easy Health” is also classified as “Easy Health”.
- Patent Document 1 describes that consumers are classified according to price sensitivity, and that a product category is determined based on cross elasticity of demand and a standard industry classification system. Patent Document 1 describes grouping products in a product category according to the type of consumer who purchases the product.
- the inventor of the present invention has conceived “a group of products similar in terms of purchasing tendency” and “a group of customers having a similar purchasing tendency” defined as follows.
- a group of products similar in terms of purchasing tendency refers to a group of products in which customers belonging to the same “group of customers having a similar purchasing tendency” show similar purchasing trends.
- the group of customers having a similar purchase tendency is a group of customers who show a similar purchase tendency with respect to products belonging to the same “group of similar products from the viewpoint of purchase tendency”.
- Groups of products that are similar in terms of purchasing trends are defined based on “Groups of customers that have similar purchasing trends”, and “Groups of customers that have similar purchasing trends” It is defined based on a group of products that are similar in terms of viewpoint. Therefore, it can be said that “a group of products similar in terms of purchasing tendency” and “a group of customers having a similar purchasing tendency” have a so-called chicken-egg relationship.
- Non-Patent Document 1 the accuracy of “a group of customers having a similar purchasing tendency” may be deteriorated.
- the work of adding features to each product is performed manually. Therefore, the accuracy of the customer cluster obtained by the method described in Non-Patent Document 1 depends on the method of adding features to the product, and the accuracy of the customer cluster may deteriorate. The same applies to the method described in Non-Patent Document 2.
- the present invention accurately defines a group of customers having a similar purchasing tendency, and relates to a similar product or a group of items related to the product in terms of purchasing tendency or a service or service similar in terms of purchasing tendency. It is an object of the present invention to provide a grouping system, a grouping method, and a grouping program that can accurately determine a group of matters to be performed.
- the grouping system has a tendency for a customer to purchase a product for each combination of a customer and a product, or for each combination of a product and a product-related matter that is a matter related to the customer and the product, based on the product purchase status of the customer.
- a purchase trend calculating means for calculating the product and a grouping means for determining a group of customers and a group of product-related items based on the trend and the distribution of the trend.
- the grouping system includes a feature amount calculation unit that calculates a feature amount of a product for each product, a product purchase record obtained for each combination of a customer and a product, a distribution of product purchase results, and a feature for each product.
- Grouping means for determining a group of customers and a group of products based on the quantity and the distribution of the feature quantity.
- the grouping system enables a customer to purchase a service for each combination of a customer and a service, or for each combination of a service-related item that is a matter related to the customer and the service, based on the customer's service purchase status.
- Purchasing tendency calculating means for calculating a tendency to Based on the trend and the distribution of the trend, a group of customers is defined, and grouping means for determining a group of services or a group of service-related items is provided.
- the grouping system includes a feature amount calculation unit that calculates a feature amount of a service for each service, a service purchase record obtained for each combination of a customer and a service, a distribution of service purchase record, and a feature for each service. And grouping means for determining a customer group and a service group based on the quantity and the distribution of the feature quantity.
- the grouping method according to the present invention enables a customer to purchase a product for each combination of a customer and a product, or for each combination of a product related item that is a matter related to the customer and the product, based on the product purchase status of the customer.
- a customer group is defined based on the trend and the distribution of the trend, and a product group or a product-related item group is defined.
- the grouping method according to the present invention calculates the feature quantity of each product, and the product purchase results obtained for each combination of customers and products, the distribution of product purchase results, the feature values for each product, and the feature values A customer group and a product group based on the distribution of the product.
- the grouping method according to the present invention enables a customer to purchase a service for each combination of a customer and a service, or for each combination of a service-related item that is a matter related to the customer and the service, based on the customer's service purchase status.
- a customer group is defined based on the trend and the distribution of the trend, and a service group or a service-related item group is defined.
- the grouping method calculates service feature values for each service, obtains service purchase results obtained for each combination of customers and services, distribution of service purchase results, feature values for each service, and feature values. And determining a group of customers and a group of services based on the distribution of the services.
- the grouping program allows a customer to store a computer for each combination of a customer and a product, or for each combination of a product and a product-related matter that is a matter related to the product. Based on the purchase trend calculation process for calculating the tendency to purchase products and the trend and distribution of the trend, the group of customers is determined and the grouping process for determining the group of products or the group of product related items is executed. It is characterized by.
- the grouping program according to the present invention includes a computer, a feature amount calculation process for calculating a feature amount of a product for each product, a product purchase record obtained for each combination of a customer and a product, a distribution of product purchase results, A grouping process for determining a group of customers and a group of products is executed based on the feature value for each product and the distribution of the feature value.
- the grouping program allows a customer to store a computer for each combination of a customer and a service, or for each combination of a service and a service-related matter that is a matter related to the customer and the service, based on the service purchase status of the customer.
- Purchasing trend calculation processing for calculating a tendency to purchase a service
- grouping processing for determining a group of customers and a group of services or a service related item based on the trend and the distribution of the trend. It is characterized by.
- the grouping program according to the present invention includes a computer, a feature amount calculation process for calculating a feature amount of a service for each service, a service purchase record obtained for each combination of a customer and a service, a distribution of service purchase record, A grouping process for determining a customer group and a service group is executed based on the feature amount for each service and the distribution of the feature amount.
- a group of customers having a similar purchasing tendency is accurately determined and related to a similar product or a group of items related to the product in terms of purchasing tendency or a service or service similar in terms of purchasing tendency.
- a group of matters to be performed can be accurately determined.
- FIG. 1 is a block diagram illustrating an example of a grouping system according to a first embodiment of this invention.
- the grouping system 1 according to the first embodiment includes a data storage unit 2, a purchase tendency calculation unit 3, and a grouping unit 4.
- the data storage means 2 is a storage device for storing purchase data indicating the customer's product purchase status.
- the data storage unit 2 stores, for example, a customer master and a product master, and stores information that associates a customer ID, a product ID, a sales price of the product, and a purchase date on each purchase date when the customer purchases the product. May be stored as
- the customer master is information indicating the attributes (age, sex, etc.) of each customer.
- the customer master is a set of information in which a customer ID is associated with each attribute of the customer.
- the product master is information indicating the attributes (product name, standard price, product category, release date, etc.) of each product.
- the product master is a set of information in which the product ID is associated with each attribute of the product.
- the customer ID is customer identification information
- the product ID is product identification information.
- the purchasing tendency calculation means 3 calculates a purchasing tendency index value for each combination of customer and product based on the purchasing data.
- the purchase tendency index value is an index value indicating a tendency of a customer to purchase a product.
- Various values can be used as purchase tendency index values. An example of the operation in which the purchase tendency calculation unit 3 calculates the purchase tendency index value will be described later.
- the purchasing trend calculation means 3 does not have to calculate a purchasing trend index value for the customer and the combination of the product.
- a customer is identified by a customer ID.
- the customer ID is represented by the symbol c.
- the product is identified by the product ID.
- the product ID is represented by the symbol i.
- the purchase trend index value calculated by the purchase trend calculation means 3 for the combination of the customer with the customer ID “c” and the product with the product ID “i” is denoted as x c, i .
- the purchase tendency index value calculated for the combination of the customer with the customer ID “2” and the product with the product ID “5” is denoted as x 2,5 .
- a customer with a customer ID “c” is described as a customer “c”.
- the product with the product ID “i” is described as the product “i”.
- the purchasing tendency calculation means 3 designates the type of distribution of the purchasing tendency index value.
- the type of distribution of the purchasing tendency index value may change depending on what value is calculated as the purchasing tendency index value.
- the purchase tendency calculation means 3 may calculate a purchase tendency index value for each combination of a customer and a product related item, not for each combination of a customer and a product.
- the product related item is not a product itself but a matter related to the product.
- a commodity related item for example, a commodity category can be cited.
- the merchandise related items are not limited to the merchandise category.
- Product-related matters are also identified by ID.
- An ID for identifying a product-related item is represented by i as in the product ID.
- the purchasing tendency index value calculated for the combination of the customer with the customer ID “c” and the product related item with the ID “i” is also denoted as x c, i .
- product-related matter with ID “i” is referred to as product-related matter “i”.
- product-related matter “i” For example, a product category with ID “1” is described as a product category “1”.
- the grouping means 4 determines a customer group and a product group.
- the grouping means 4 determines a customer group and a product related item group. For example, when the purchase tendency calculation means 3 calculates the purchase tendency index value xc, i for each combination of a customer and a product category, the grouping means 4 determines a customer group and a product category group.
- the purchasing tendency calculation means 3 may calculate any purchasing tendency index value in the following examples.
- the purchase tendency calculation unit 3 may calculate a value other than the values exemplified below as the purchase tendency index value.
- the natural logarithm of an arbitrary variable v is denoted as ln (v).
- Example 1 the purchase tendency calculation means 3 calculates the degree of change in the purchase amount of a product in accordance with a price reduction (a discount rate is acceptable) of the product as a purchase trend index value. For example, the purchase tendency calculation means 3 calculates price elasticity as a purchase tendency index value.
- the purchase tendency calculation means 3 creates, for example, information in which the purchase amount is associated with the combination of the customer ID, the product ID, and the actual sales price of the product based on the purchase data. At this time, even if the customer IDs are common and the product IDs are common, if the selling price is different, the purchase tendency calculation means 3 is set to another group. That is, the purchase tendency calculation means 3 specifies the number of purchases for each actual selling price of the product for the combination of the customer ID and the product ID.
- the purchase tendency calculation means 3 refers to the above information created based on the purchase data, and performs a regression analysis using ln (volume c, i ) as an objective variable and ln (price c, i ) as an explanatory variable. By this regression analysis, w c, i in the following equation (1) is obtained.
- the purchase tendency calculation means 3 calculates w c, i in the equation (1) as a purchase tendency index value x c, i .
- purchasing habits index value x c, as i may be the absolute value of the above w c, i. This value is price elasticity.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- Example 2 the purchase trend calculation means 3 calculates a value indicating the degree of days from the advertisement date of a product until the customer purchases the product as a purchase trend index value.
- the data storage means 2 also stores information in which the product ID is associated with the advertisement date of the product.
- the purchase tendency calculation means 3 uses, as purchase data, information that associates a customer ID, a product ID, and the number of days elapsed since the most recent advertisement date on each purchase date when the customer purchased the product. Create based on.
- the number of days that have elapsed since the advertisement date of the product on each purchase date when the customer “c” purchased the product “i” is referred to as an elapsed days variable day c, i .
- the purchase quantity that the customer “c” purchased the product “i” on each purchase date is described as a purchase quantity variable volume c, i .
- the purchase tendency calculation means 3 refers to the above information created based on the purchase data, and performs regression analysis using ln (volume c, i ) as an objective variable and ln (day c, i ) as an explanatory variable. W c, i in the equation (2) shown in FIG.
- the purchase tendency calculation means 3 calculates w c, i in the equation (2) as a purchase tendency index value x c, i .
- the purchase tendency calculating means 3 may obtain the reciprocal of the absolute value of w c, i as the purchase tendency index value x c, i . This value can be said to be the advertisement effective lifetime of the product “i” for the customer “c”.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- Example 3 the purchase trend calculation means 3 calculates a value indicating the degree of days from when a new product is released until the customer purchases the product as a purchase trend index value.
- the purchase trend calculation means 3 uses the purchase data as the information that associates the customer ID, the product ID, and the number of days elapsed from the release date of the product on each purchase date when the customer purchased the product. Create based on.
- the elapsed days from the product release date on each purchase date when the customer “c” purchased the product “i” is denoted as an elapsed day variable day c, i .
- the purchase quantity that the customer “c” purchased the product “i” on each purchase date is described as a purchase quantity variable volume c, i .
- the purchase tendency calculation means 3 refers to the above information created based on the purchase data, and performs regression analysis using ln (volume c, i ) as an objective variable and ln (day c, i ) as an explanatory variable. W c, i in the equation (3) shown in FIG.
- the purchase tendency calculation means 3 calculates w c, i in the equation (3) as a purchase tendency index value x c, i .
- the purchase tendency calculating means 3 may obtain the reciprocal of the absolute value of w c, i as the purchase tendency index value x c, i . This value can be said to be the new product sensitivity life.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- the purchase tendency calculation means 3 may designate a Poisson distribution.
- Example 4 the purchase trend calculation means 3 calculates a value indicating the degree of days from when a product is displayed at a store until the customer purchases the product as a purchase trend index value.
- the data storage means 2 also stores information in which the product ID is associated with the display date of the product.
- the purchase tendency calculation means 3 Based on the purchase data, the purchase tendency calculation means 3, for example, associates the customer ID, the product ID, and the number of days elapsed from the display date of the product on each purchase date when the customer purchased the product. create.
- the number of days elapsed from the display date of the product on each purchase date when the customer “c” purchased the product “i” is referred to as an elapsed days variable day c, i .
- the purchase quantity that the customer “c” purchased the product “i” on each purchase date is described as a purchase quantity variable volume c, i .
- the purchase tendency calculation means 3 refers to the above information created based on the purchase data, and performs regression analysis using ln (volume c, i ) as an objective variable and ln (day c, i ) as an explanatory variable. Wc, i in the equation (4) shown in FIG.
- the purchase tendency calculation means 3 calculates w c, i in the equation (4) as a purchase tendency index value x c, i .
- the purchase tendency calculating means 3 may obtain the reciprocal of the absolute value of w c, i as the purchase tendency index value x c, i .
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- the purchasing tendency calculation means 3 associates the customer ID, the product ID, and the number of days elapsed from the display date of the product. Information may be created.
- Example 5 the purchase trend calculation means 3 calculates a value indicating the degree of days from when a customer purchases a product to the next purchase of the product as a purchase trend index value.
- the purchase tendency calculation means 3 creates, for example, information in which a customer ID, a product ID, and an average purchase interval of products are associated with each other based on purchase data.
- the average purchase interval corresponding to the combination of the customer ID and the product ID is set as a purchase tendency index value xc, i .
- the purchase tendency calculation means 3 calculates the elapsed days from the date when the customer “c” purchased the product “i” to the next date when the same customer “c” purchased the same product “i”. Then, the average value of the elapsed days is calculated as the purchase tendency index value xc, i .
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- the purchasing tendency calculation means 3 calculates a purchasing tendency index value x c, i for each combination of customer and product.
- Example 6 the purchase tendency calculation means 3 calculates the degree of customer's commitment (in other words, the degree of attachment) to individual products in the product category as a purchase tendency index value.
- the purchasing tendency calculation means 3 calculates, for example, a Manton coefficient or Gini coefficient as the purchasing tendency index value xc, i .
- the purchase tendency calculation means 3 obtains the purchase share in the product category “i” by the customer “c” for the product belonging to the product category “i”, and the product category Calculate the sum of the squares of the purchase shares obtained for each product belonging to “i”.
- the purchasing tendency calculation means 3 uses this value (Hafender coefficient) as the purchasing tendency index value xc, i .
- the purchasing tendency calculation means 3 calculates a purchasing tendency index value xc, i for each combination of customer and product category.
- x c, i is a small value indicates that the customer “c” buys various products of the product category “i” and seeks diversity in the products.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- product DNA may be applied as a product-related item instead of the product category.
- Example 7 the purchase tendency calculation means 3 calculates the degree of customer's commitment (in other words, the degree of attachment) to the manufacturer of individual products in the product category as a purchase tendency index value.
- the purchase tendency calculation means 3 calculates, for example, a Human Capital Average coefficient or Gini coefficient as the purchase tendency index value xc, i .
- a Human Capital Average coefficient or Gini coefficient For example, when calculating the Manton coefficient, for each manufacturer belonging to a certain product category “i”, the purchase share in the product category “i” by the customer “c” is obtained and belongs to the product category “i”. Calculate the sum of the squares of the purchase shares obtained for each manufacturer.
- the purchasing tendency calculation means 3 uses this value (Hafender coefficient) as the purchasing tendency index value xc, i .
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- Example 7 a product brand or the like may be applied instead of the manufacturer.
- the purchasing tendency calculation means 3 calculates a purchasing tendency index value x c, i for each combination of a customer and a product related item (product category in the above example).
- the grouping unit 4 uses the purchase trend index values x c, i calculated by the purchase trend calculation unit 3 and the distribution types of the specified purchase trend index values x c, i to determine the customer groups and products. Define a group.
- the grouping means 4 may determine a group of customers and a group of merchandise related items (for example, merchandise category). In any case, the operation of the grouping means 4 is the same. In the following description, a case where the grouping means 4 determines a customer group and a product group will be described as an example.
- one customer belongs to only one group
- one product belongs to only one group.
- grouping A case where grouping is performed will be described as an example. In this way, defining a group so that one element belongs to only one group is called clustering. A group obtained by clustering is called a cluster. Further, simultaneous clustering for a plurality of objects (customers and products in this example) is called Co-clustering. That is, in the following description, a case where the grouping unit 4 executes Co-clustering will be described as an example.
- FIG. 2 is a diagram schematically showing a customer ID and a product ID before execution of Co-clustering.
- a state in which customer IDs are sequentially arranged in the vertical axis direction and product IDs are sequentially arranged in the horizontal axis direction is illustrated.
- Each purchase tendency index value x c, i calculated by the purchase tendency calculation means 3 corresponds to a combination of one customer ID and one product ID.
- x 2,1 shown in FIG. 2 corresponds to a combination of customer ID “2” and product ID “1”.
- the purchase tendency index value x c, i does not have to be calculated for the combination of the customer and the product.
- FIG. 3 is an explanatory diagram schematically showing an example of a customer group (customer cluster) and a product group (product cluster) defined by the grouping means 4.
- the grouping means 4 determines a plurality of product clusters and a plurality of customer clusters.
- the number of product clusters and the number of customer clusters may be set to fixed values or may not be limited to fixed values.
- the commodity cluster ID is “a”
- the commodity cluster is referred to as commodity cluster “a”.
- the customer cluster ID is “b”
- the customer cluster is described as a customer cluster “b”.
- the products “6”, “8”, and “10” belong to the product cluster “9” determined by the grouping means 4. Further, it is assumed that customers “2”, “5”, and “9” belong to the customer cluster “3” determined by the grouping means 4.
- the number of products (product ID) belonging to the product cluster and the number of customers (customer ID) belonging to the customer cluster are not particularly limited.
- a combination of one product cluster and one customer cluster corresponds to a purchase tendency index value xc, i corresponding to the combination of the product belonging to the product cluster and the customer belonging to the customer cluster.
- x 2,6 , x 5,6 and the like correspond to the combination of the product cluster “9” and the customer cluster “3”.
- the distribution parameters of the purchasing tendency index values xc, i in the specified distribution type are also determined for each combination of one product cluster and one customer cluster. become. For example, distribution parameters such as x 2,6 , x 5,6 corresponding to the combination of the product cluster “9” and the customer cluster “3” are also determined.
- distribution parameters such as x 2,6 , x 5,6 corresponding to the combination of the product cluster “9” and the customer cluster “3” are also determined.
- the parameters corresponding to the combination of the product cluster “9” and the customer cluster “3” are represented as ⁇ 3,9 .
- Examples of distribution parameters of xc, i include, for example, an average value and variance.
- ⁇ 3 , 9 and the like may be expressed as a vector having an average value, variance, and the like as elements, for example.
- FIG. 3 is a modification of FIG. 2 so that product IDs belonging to the same product cluster are continuously arranged, and customer IDs belonging to the same customer cluster are continuously arranged. it can.
- the grouping means 4 has, for example, a plurality of sets of likelihoods obtained by combining customer clusters (customer groups), product clusters (product groups), and parameters of distribution of purchase tendency index values xc, i. Each customer cluster and each product cluster are defined so as to be maximized. This likelihood is expressed by the following equation (5).
- the grouping unit 4 may obtain a customer cluster and a product cluster by performing processing for maximizing the lower limit of the marginal likelihood obtained by marginalizing ⁇ shown in Expression (5), or processing similar thereto.
- C is a set of customer IDs
- I is a set of product IDs.
- C is a set of customers and I is a set of products.
- Z c represents the customer cluster to which the customer ID “c” belongs.
- Z i represents the product cluster to which the product ID “i” belongs.
- the grouping means 4 executes Co-clustering so that one customer ID belongs to only one customer cluster and one product ID belongs to only one product cluster.
- ⁇ is a distribution parameter designated by the purchasing tendency calculation means 3.
- ⁇ is a parameter of distribution of x c, i corresponding to a combination of one customer cluster indicated by Z c and one commodity cluster indicated by Z i .
- Z c indicates the customer cluster “3” shown in FIG. 3
- Z i indicates the product cluster “9” shown in FIG. 3
- ⁇ 3,9 is assigned to the set of Z c and Z i .
- ⁇ , Z c , Z i ) represents the probability (probability) of x c, i .
- the value of Expression (5) is the likelihood as a whole of a plurality of sets obtained by combining the customer cluster, the product cluster, and the parameters of the distribution of xc, i corresponding to the two clusters.
- the grouping means 4 determines a customer cluster and a product cluster by determining a plurality of sets of customer clusters, product clusters and parameters so that the value of the equation (5) is maximized. Good.
- the grouping unit 4 calculates the posterior distribution of ⁇ shown in Equation (5).
- the grouping unit 4 uses the EM (Expectation-Maximization) method to set the combination of the customer cluster, the product cluster, and the parameters so that the value of Expression (5) becomes maximum. It suffices to determine more than one. If it is Bayesian estimation, the grouping means 4 calculates
- the Gibbs sampling method is one of MCMC methods (Markov Chain Monte Carlo algorithm).
- the purchase tendency calculation means 3 and the grouping means 4 are realized by a CPU of a computer, for example.
- the CPU only has to read the grouping program from a program recording medium such as a computer program storage device (not shown in FIG. 1) and operate as the purchase tendency calculating means 3 and the grouping means 4 according to the grouping program.
- Each means may be realized by separate hardware.
- the grouping system may have a configuration in which two or more physically separated devices are connected by wire or wirelessly. This also applies to embodiments described later.
- FIG. 4 is a flowchart showing an example of processing progress of the first embodiment of the present invention.
- the purchasing tendency calculation means 3 calculates a purchasing tendency index value xc, i for each combination of customer and product based on the purchase data (step S1). In step S1, the purchasing tendency calculation means 3 designates the distribution type of the purchasing tendency index value xc, i .
- step S1 the grouping means 4 determines a plurality of sets of customer clusters, product clusters, and xc, i distribution parameters so that the value of the equation (5) is maximized, whereby the customer clusters and products.
- a cluster is determined (step S2). This distribution is designated by the purchasing tendency calculation means 3.
- a human does not give any feature to a product.
- Xc, i obtained in step S1 is objective data obtained from purchase data.
- the grouping means 4 defines a group of customers and a group of products (in the above example, a customer cluster and a product cluster), so that a group of customers having a similar purchasing tendency can be accurately defined. It is possible to accurately determine a group of products that are similar in terms of purchasing tendency.
- the grouping means 4 is configured so that the value of the expression (5) is maximized.
- the customer cluster and the product category (product related item) cluster may be determined.
- the grouping unit 4 may perform Co-clustering in step S2 using the product category ID as “i” in the equation (5) instead of the product ID.
- the other points are the same as the operation of the grouping means 4 already described. In this case, it is possible to accurately determine a group of customers having a similar purchasing tendency and to accurately determine a group of similar product categories from the viewpoint of purchasing tendency.
- the grouping means 4 allows the group of customers while allowing one customer (customer ID) to belong to a plurality of groups and one product (product ID) to belong to a plurality of groups. And product groups may be defined. Also in this case, the grouping means 4 may determine the customer group and the product group so that the value of the expression (5) is maximized. Even in this case, the grouping means 4 may use the Gibbs sampling method, the EM method, or the variational Bayes method as a method of determining the customer group and the product group so that the value of the expression (5) is maximized. . The same applies to a group of product-related items such as product categories.
- the grouping unit 4 determines a customer group and a product group while allowing one customer (customer ID) to belong to a plurality of groups and one product (product ID) to belong to a plurality of groups.
- the data analyst can find hidden features (features that are difficult to find directly) of the customer or the product by referring to the group of the customer and the group of the product.
- the purchase tendency index value x c, i for the product i of the customer c is modeled as f (Z T i AZ c ) using the function f.
- This function f is a logistic function, Poisson distribution, Gaussian distribution, or the like.
- the function f may be determined by the purchase tendency index value xc , i . If the purchasing tendency index value x c, i is a value of 0 or 1, the function f is a logistic function. If the purchasing tendency index value x c, i is a number, the function f has a Poisson distribution. If the purchase tendency index value x c, i is a real number, the function f has a Gaussian distribution.
- the probability distribution shown as a specific example in the embodiment corresponds to this function.
- the purchase tendency calculation unit 3 specifies the type of distribution of the purchase tendency index value xc, i .
- Buying habits index value x c depending on whether using what value as i, buying habits index value x c, the type of distribution of the i may be determined in advance.
- the purchasing tendency calculation means 3 does not have to specify the type of distribution of the purchasing tendency index value xc, i .
- the grouping unit 4 may determine a group of customers and a group of products or product-related items by using distribution parameters of purchase tendency index values xc, i in a predetermined distribution type.
- FIG. FIG. 5 is a block diagram illustrating an example of a grouping system according to the second embodiment of this invention.
- the grouping system 11 according to the second embodiment includes a data storage unit 12, a feature amount calculation unit 13, and a grouping unit 14.
- the data storage means 12 is a storage device for storing purchase data indicating the customer's product purchase status.
- the customer master and the product master may be stored, and information in which the customer ID, the product ID, the price of the product, and the purchase date are associated with each purchase date when the customer purchased the product may be stored as purchase data.
- the feature amount calculation unit 13 calculates the relative price of a product as the feature amount of the product using the standard price in the product master will be described as an example.
- the number of purchases is used as an index value indicating a product purchase record (hereinafter referred to as a purchase record index value).
- the feature amount calculation means 13 calculates the feature amount of the product for each product. This feature amount is a feature amount of the product itself and does not depend on the customer. In this example, the case where the feature amount calculation unit 13 calculates the relative price of each product as the feature amount of the product will be described as an example. In this case, the feature amount calculation unit 13 is based on the standard price of each product stored in the data storage unit 12 and the standard price (hereinafter, referred to as i av ) of the standard price of each product. The standard deviation (hereinafter referred to as i dev ) is calculated. The standard price of the product “i” stored in the data storage unit 12 is set to s i . Also, let the relative price of the product “i” be r i . The feature quantity calculation means 13 calculates the relative price of each product by performing the following equation (6) for each product.
- the feature amount calculation means 13 designates the distribution type of the feature amount of the product.
- the feature quantity calculation means 13 designates, for example, a standard normal distribution as the distribution type of the feature quantity (relative price).
- the feature amount calculation unit 13 may specify a Gaussian distribution.
- the grouping unit 14 refers to the purchase data stored in the data storage unit 12 and calculates a purchase record index value for each combination of customer and product.
- a purchase performance index value calculated for each combination of a customer and a product is represented by a symbol uc , i .
- the grouping means 14 employs a Poisson distribution as a distribution of purchase performance index values (number of purchases).
- the grouping means 14 determines the number of purchases derived for each combination of customer and product, the type of distribution of the number of purchases (Poisson distribution in this example), the relative price for each product, and the distribution of the relative price. Define customer groups and product groups based on type.
- grouping is performed so that one customer (in other words, one customer ID) belongs to only one cluster, and one product (in other words, one product ID) belongs to only one cluster.
- a case where the means 14 executes Co-clustering will be described as an example.
- the grouping unit 14 sets each customer so that the likelihood of a plurality of sets obtained by combining the customer cluster, the product cluster, the distribution parameter of the number of purchases, and the distribution parameter of the relative price of the product is maximized. Define clusters and each product cluster. This likelihood is expressed by the following equation (7).
- C is a set of customer IDs
- I is a set of product IDs
- Z c represents the customer cluster to which the customer ID “c” belongs.
- Z i represents the product cluster to which the product ID “i” belongs. This point is the same as in the first embodiment.
- ⁇ ′ is a parameter of the distribution of the number of purchases u c, i corresponding to the combination of one customer cluster indicated by Z c and one product cluster indicated by Z i .
- ⁇ ′ may be represented by a vector, for example.
- Equation (7) p (u c, i
- ⁇ is a parameter of the distribution designated by the feature amount calculation unit 13.
- ⁇ is a parameter of the relative price distribution of products belonging to the product cluster Z i .
- ⁇ , Z i ) represents the probability of r i .
- the value of Expression (7) is the distribution of the customer cluster, the product cluster , the distribution parameter ⁇ ′ of uc, i corresponding to the two clusters, and the relative price (feature value) of the product belonging to the product cluster. It can be said that it is the likelihood as a whole of a plurality of sets obtained by combining the parameter ⁇ .
- the grouping unit 14 may determine the customer cluster and the product cluster by determining a plurality of sets of customer clusters, product clusters, and parameters ⁇ ′ and ⁇ so that the value of the expression (7) is maximized.
- the grouping means 14 uses the Gibbs sampling method, the EM method, or the variational Bayes method, for example, so as to maximize the value of the expression (7), the customer cluster, the product cluster, and the parameters ⁇ ′, ⁇ What is necessary is just to define two or more sets.
- the feature quantity calculation means 13 and the grouping means 14 are realized by a CPU of a computer, for example.
- the CPU may read the grouping program from a program recording medium such as a computer program storage device (not shown in FIG. 5), and operate as the feature amount calculation unit 13 and the grouping unit 14 in accordance with the grouping program.
- Each means may be realized by separate hardware.
- FIG. 6 is a flowchart showing an example of processing progress of the second embodiment of the present invention.
- Feature calculating unit 13 for example, based on the product master calculates the relative price r i Product for each product (step S11).
- the feature quantity calculation means 13 may calculate the relative price of each product by calculating the formula (6) for each product. Further, in step S11, the feature amount calculating unit 13 specifies the type of the distribution of the relative price r i.
- the grouping unit 14 refers to the purchase data and derives the number of items purchased by the customer (number of purchases u c, i ) for each combination of the customer and the item (step S12). Note that the number of purchases uc , i may be stored in the data storage unit 12 for each combination of customer and product. In that case, the grouping means 14 may read each purchase quantity u c, i from the data storage means 12.
- the grouping unit 14 may determine the customer cluster and the product cluster by determining a plurality of sets of the customer cluster, the product cluster, and the parameters ⁇ ′ and ⁇ so that the value of Expression (7) is maximized. (Step S13).
- a human does not give any feature to a product.
- ri, uc, and i in this embodiment are objective data obtained from the product master and purchase data. Using such data, the grouping means 14 determines customer clusters and product clusters. Accordingly, it is possible to accurately determine a group of customers having a similar purchasing tendency and to accurately determine a group of similar products from the viewpoint of purchasing tendency.
- the grouping means 14 determines a customer group and a product group while allowing one customer (customer ID) to belong to a plurality of groups and one product (product ID) to belong to a plurality of groups. Also good. Also in this case, the grouping means 14 can determine the customer cluster and the product cluster by determining a plurality of combinations of the customer cluster, the product cluster, and the parameters ⁇ ′ and ⁇ so that the value of the expression (7) is maximized. That's fine. Even in this case, the grouping means 14 may use the Gibbs sampling method, the EM method, or the variational Bayes method as a method of determining the customer group and the product group so that the value of the expression (7) is maximized. .
- the grouping means 14 defines a group of customers and a group of products while allowing one customer (customer ID) to belong to a plurality of groups and one product (product ID) to belong to a plurality of groups.
- the data analyst can find hidden features (features that are difficult to find directly) of the customer or the product by referring to the group of the customer and the group of the product.
- the feature amount calculation unit 13 specifies the type of distribution of the feature amount of the product.
- the type of distribution of the feature amount of the product may be determined in advance.
- the feature amount calculation unit 13 does not have to specify the type of distribution of the feature amount of the product.
- the grouping unit 14 may determine the customer group and the product group using the distribution parameter of the feature amount of the product in the predetermined distribution type.
- FIG. 7 is a schematic diagram showing an example of a customer master stored in the data storage unit 2 of the first embodiment and the data storage unit 12 of the second embodiment.
- FIG. 7 illustrates a case where the customer ID is associated with the customer's age and sex for each customer ID.
- FIG. 8 is a schematic diagram illustrating an example of a product master stored in the data storage unit 2 of the first embodiment and the data storage unit 12 of the second embodiment.
- FIG. 8 illustrates a case where a product ID is associated with a product name, a standard price, a product category, and a release date for each product ID.
- FIG. 9 is a schematic diagram showing an example of purchase data stored in the data storage means 2 of the first embodiment and the data storage means 12 of the second embodiment.
- FIG. 9 illustrates purchase data in which a customer ID, a product ID, an actual sales price of a product, and a purchase date are associated with each purchase date when the customer purchases the product. Note that the data in one row shown in FIG. 9 indicates that the customer has purchased one product.
- FIGS. 7 to 9 are examples of data stored in the data storage means 2 and 12.
- FIG. Specific numerical values shown in the specific examples described below are not necessarily based on the data shown in FIGS.
- Specific example 1 shown below shows a specific example of the feature amount of the product in the second embodiment.
- the feature quantity calculation means 13 calculates the feature quantity of each product in the category for each product category. In this example, the case where the feature amount calculation means 13 calculates a relative price as the feature amount of the product is taken as an example.
- the feature amount calculating means 13 selects a product category one by one, and calculates an average value i av of the standard price of each product in the selected product category and a standard deviation i dev of the standard price of each product.
- FIG. 10 is a diagram illustrating examples of average values and standard deviations calculated for two product categories of “confectionery bread” and “bread bread”, respectively.
- the feature quantity calculation means 13 selects a product category one by one, and calculates the relative price r i of each product in the selected product category using the above-described equation (6).
- i av and i dev in the calculation of Equation (6) are the average value i av and the standard deviation i dev calculated for the selected product category.
- FIG. 11 is a diagram illustrating an example of the relative price calculated for each product.
- FIG. 11 shows an example of product names and relative prices for each product ID.
- the feature quantity calculation unit 13 designates, for example, a Gaussian distribution as the type of distribution of the relative price r i calculated as described above.
- the feature amount calculation unit 13 may specify a standard normal distribution as the distribution type of the relative price r i .
- the grouping means 14 refers to the purchase data, calculates the number of items that the customer has purchased the item (the number of purchases) for each combination of the customer ID and the item ID, and uses the number of purchases as the purchase performance index value. Let u c, i . Furthermore, the grouping means 14 employs a Poisson distribution as a distribution of purchase performance index values u c, i (hereinafter, the number of purchases u c, i ).
- the grouping unit 14 determines the customer ID and the product ID based on the relative price r i of each product and the type of distribution thereof, and the number of purchases u c, i derived for each combination of the customer ID and product ID and the type of distribution thereof. Group. Since this process has already been described, the description thereof is omitted here.
- Specific example 2 shows a specific example of example 1 described in the first embodiment.
- the purchase tendency calculation means 3 specifies the number of purchases (purchase amount) for each actual sales price of the product for the combination of the customer ID and the product ID based on the purchase data. As a result, as illustrated in FIG. 12, information in which the number of purchases is associated with the combination of the customer ID, the product ID, and the actual sales price of the product is obtained.
- the purchasing tendency calculation means 3 performs a regression analysis with ln (volume c, i ) as an objective variable and ln (price c, i ) as an explanatory variable.
- volume c, i is a purchase quantity variable
- price c, i is a price variable.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- the grouping unit 4 groups customer IDs and product IDs based on the purchase tendency index value xc, i and the type of distribution thereof. Since this process has already been described, the description thereof is omitted here. This also applies to other specific examples described below.
- Specific Example 3 shows a specific example of Example 2 described in the first embodiment.
- FIG. 13 is a diagram illustrating an example of information in which a product ID and an advertisement date are associated with each other.
- the advertisement may be a commercial message for television broadcasting or radio broadcasting.
- the information illustrated in FIG. 13 may reflect the results of advertisements broadcasted together in conjunction with a program.
- the information illustrated in FIG. 13 may be created based on an advertisement (for example, online advertisement) that is individually executed for each customer.
- the purchase tendency calculation means 3 uses, as purchase data, information that associates a customer ID, a product ID, and the number of days elapsed since the most recent advertisement date on each purchase date when the customer purchased the product. Create based on.
- FIG. 14 shows an example of information in which the customer ID, the product ID, and the number of days elapsed from the advertisement date are associated with each other.
- the purchase tendency calculation means 3 may treat the number of days elapsed from the advertisement date as a sufficiently large value or a missing value. “Inf.” Shown in FIG. 14 means a sufficiently large value.
- the purchase tendency calculation means 3 calculates the advertisement effective lifetime T c, i of the product for the customer for each combination of the customer “c” and the product “i”.
- the number of purchases in the number of days c, i since the advertisement date is proportional to exp ( ⁇ day c, i / T c, i )
- the probability that the customer “c” purchases the product “i” in the days c, i since the advertisement date is proportional to exp ( ⁇ day c, i / T c, i ).
- the former case is shown.
- the purchasing tendency calculation means 3 performs a regression analysis using ln (volume c, i ) as an objective variable and ln (day c, i ) as an explanatory variable.
- volume c, i is a purchase quantity variable
- day c, i is an elapsed day variable.
- ⁇ 0.2 is obtained as the value of the coefficient w c, i in the equation (2).
- the purchasing tendency calculation means 3 similarly calculates the purchasing tendency index value xc, i for each other combination of the customer ID and the product ID.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- the purchase tendency calculation means 3 may use the number of days elapsed from the display date of the product instead of the number of days elapsed from the advertisement date.
- the purchase tendency calculation means 3 may specify, for example, a normal distribution or a Poisson distribution as the distribution type of the purchase tendency index value xc, i .
- Specific example 4 shows an example in which the purchase tendency index value x c, i is obtained using the number of days elapsed from the sale date of the product.
- the purchase tendency calculation means 3 uses information on the purchase date that associates the customer ID, the product ID, and the number of days elapsed from the release date of the product on each purchase date when the customer purchased the product. create.
- FIG. 15 shows an example of information in which the customer ID, the product ID, and the number of days elapsed from the release date are associated with each other.
- the purchase tendency calculation means 3 calculates a new product sensitivity life T c, i of the product for the customer for each combination of the customer “c” and the product “i”.
- the probability that the customer “c” purchases the product “i” in the number of days c, i since the product release date is exp ( ⁇ day c, i / Assume that it is proportional to T c, i ).
- the purchase tendency calculating means 3 sets the new product sensitivity lifetime T 01,11 that maximizes the likelihood of sensitivity to the product “11” of the customer “01” as the advertisement effective lifetime T 01,11 in the above specific example. It can be obtained similarly.
- the purchase trend calculation means 3 sets the new product sensitivity life T 01,11 as the purchase trend index value x 01,11 for the combination of the customer ID “01” and the product ID “11”.
- the purchasing tendency calculation means 3 similarly calculates the purchasing tendency index value xc, i for each other combination of the customer ID and the product ID.
- the purchase tendency calculation means 3 designates, for example, a Poisson distribution as the distribution type of the purchase tendency index value xc, i .
- the purchase tendency calculation means 3 may specify a normal distribution.
- Specific example 5 shows a specific example of example 5 described in the first embodiment.
- the purchasing tendency calculation means 3 calculates the elapsed days from the date when the customer “c” purchased the product “i” to the next date when the same customer “c” purchased the same product “i”, respectively.
- the average value of the elapsed days is calculated as the purchase tendency index value xc, i .
- the average purchase interval corresponds to the optimal parameter of the distribution when the purchase interval follows the Poisson distribution.
- the purchase tendency calculation means 3 calculates an average purchase interval (purchase tendency index value x c, i ) for each combination of customer ID and product ID. And the purchase tendency calculation means 3 produces the information which matched customer ID, goods ID, and an average purchase space
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- Specific Example 6 shows a specific example of Example 6 described in the first embodiment.
- the product master illustrated in FIG. 17 is stored in the data storage unit 2 (see FIG. 1).
- a product ID, a product name, a product category, and a product manufacturer are associated with each product ID.
- the purchase data illustrated in FIG. 18 is stored in the data storage unit 2 (see FIG. 1).
- the number of purchases (the number of purchases) in which the customer identified by the customer ID purchases the product identified by the product ID is associated with the combination of the customer ID and the product ID. .
- the purchasing tendency calculation means 3 calculates the customer's commitment (attachment level) to the individual product in the product category.
- the purchasing tendency calculation means 3 calculates a Skinder coefficient as the degree of sticking and sets the value as the purchasing tendency index value xc, i .
- the purchase tendency calculating means 3 calculates the degree of sticking regarding the combination of the customer ID “01” and the product category “confectionery bread” as an example.
- the customer “01” has one anpan (product ID “11”) and a high-quality bread (product ID “12”) as products belonging to the product category “confectionery bread”.
- Two pieces of curry bread (product ID “15”) and one piece of melon bread (product ID “16”) are purchased. Therefore, the purchase shares of “an”, “high-quality”, “curry” and “melon” belonging to “sweet bread” by the customer “01” are “1/6”, “1/3”, “1/3” and “1”, respectively. / 6 ". Therefore, the purchasing tendency calculation means 3 calculates 0.28 as the sticking degree (in this example, the Gross coefficient) in the combination of the customer ID “01” and the product category “confectionery bread” by the following calculation.
- the purchase tendency calculation means 3 calculates the sticking degree regarding the combination of the customer ID “02” and the product category “confectionery bread”.
- the customer “02” has two anpan (product ID “11”), three high-quality bread (product ID “12”), and curry bread (product ID “product ID”) as products belonging to the product category “confectionery bread”. 18 ′′) is purchased (see FIGS. 17 and 18). Therefore, the purchase shares of the bread, high-quality bread and spicy curry bread belonging to the “confectionery bread” by the customer “02” are “1/3”, “1/2” and “1/6”, respectively. Therefore, the purchasing tendency calculation means 3 calculates 0.39 as the sticking degree in the combination of the customer ID “02” and the product category “confectionery bread” by the following calculation.
- the purchase tendency calculation means 3 calculates the sticking degree for the combination of the customer ID “03” and the product category “confectionery bread”.
- the customer “03” is a product belonging to the product category “confectionery bread”, two high-quality breads (product ID “12”), eight melon breads (product ID “16”), and yakisoba bread (product ID “17”). ) Is purchased (see FIGS. 17 and 18). Accordingly, the purchase shares of high-quality bread, melon bread and yakisoba bread belonging to “confectionery bread” by customer “03” are “2/11”, “8/11” and “1/11”, respectively. Accordingly, the purchasing tendency calculation means 3 calculates 0.57 as the sticking degree in the combination of the customer ID “03” and the product category “confectionery bread” by the following calculation.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- FIG. 19 is a diagram illustrating an example of the sticking degree (x c, i ) for each combination of the customer ID and the product category.
- the customer ID “01” and the customer ID “02” are grouped into the same customer group (A) by grouping by the grouping means 4, and the customer ID “03” is set as another customer group (B). )).
- “confectionery bread” and “rice ball” are considered to be grouped into the same product category group. If such a grouping result is obtained, customer group A tends to purchase various products for “confectionery bread” and “rice ball”, and tends to select their favorite products for “bread”. Can be analyzed. Further, the customer group B can analyze that there is a tendency to purchase specific products not only for “bread bread” but also for “confectionery bread” and “rice ball”.
- Specific Example 7 shows a specific example of Example 7 described in the first embodiment.
- the purchase tendency calculation means 3 calculates the degree of customer's commitment (attachment) to the manufacturer of the product in the product category.
- the purchasing tendency calculation means 3 calculates a Skinder coefficient as the degree of sticking and sets the value as the purchasing tendency index value xc, i .
- the purchase tendency calculation means 3 refers to the product master and the purchase data, for example, for the product category “confectionery bread”, the customer ID, the manufacturer of the sweet bread purchased by the customer, and the number of purchases of the sweet bread of the manufacturer. Identify relationships. Then, it is assumed that the result shown in FIG. 20 is obtained as the relationship between the customer ID, the confectionery bread maker, and the number of confectionery bread purchases by the maker.
- the purchasing tendency calculation means 3 calculates a sticking degree (in this example, a Mader coefficient) 0.5 in the combination of the customer ID “01” and the product category “confectionery bread” by the following calculation.
- the purchase tendency calculation means 3 calculates a sticking degree (a Mamony coefficient in this example) 0.72 in the combination of the customer ID “02” and the product category “confectionery bread” by the following calculation.
- the purchase tendency calculation means 3 calculates a sticking degree (a Mamony coefficient in this example) 0.70 in the combination of the customer ID “03” and the product category “confectionery bread” by the following calculation.
- the purchase tendency calculation means 3 calculates the sticking degree (x c, i ) for each combination of the customer ID and the product category.
- the purchasing tendency calculation means 3 designates, for example, a normal distribution as the type of distribution of the purchasing tendency index value xc, i .
- FIG. 21 is a schematic block diagram showing a configuration example of a computer according to each embodiment of the present invention.
- the computer 1000 includes a CPU 1001, a main storage device 1002, an auxiliary storage device 1003, and an interface 1004.
- the grouping system of each embodiment is mounted on the computer 1000.
- the operations of the grouping system are stored in the auxiliary storage device 1003 in the form of a program (grouping program).
- the CPU 1001 reads out the program from the auxiliary storage device 1003, develops it in the main storage device 1002, and executes the above processing according to the program.
- the auxiliary storage device 1003 is an example of a tangible medium that is not temporary.
- Other examples of the non-temporary tangible medium include a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, and a semiconductor memory connected via the interface 1004.
- this program is distributed to the computer 1000 via a communication line, the computer 1000 that has received the distribution may develop the program in the main storage device 1002 and execute the above processing.
- the program may be for realizing a part of the above-described processing.
- the program may be a differential program that realizes the above-described processing in combination with another program already stored in the auxiliary storage device 1003.
- FIG. 22 is a block diagram showing an example of the outline of the present invention.
- the grouping system 1 includes purchase tendency calculation means 3 and grouping means 4.
- Purchasing tendency calculation means 3 uses the customer's product purchase status for each combination of the customer and the product, or for each combination of the customer and the product related matter (for example, product category) that is a matter related to the product. Calculates the tendency to purchase goods.
- the grouping unit 4 determines a group of customers and a group of products or a product related item based on the trend and the distribution of the trend.
- the purchase tendency calculation means 3 is based on the purchase data indicating the customer's product purchase status, for each combination of the customer and the product, or for each combination of the product related items that are items related to the customer and the product.
- a purchase tendency index value which is an index value indicating the tendency of the customer to purchase the product.
- the grouping means 4 determines a group of customers and a group of products or a group of product related items based on the purchase trend index value and the distribution parameter of the purchase trend index value.
- the grouping system 1 may determine a group of services or a group of service-related matters (for example, service categories).
- FIG. 23 is a block diagram showing another example of the outline of the present invention.
- the grouping system 11 includes a feature amount calculation unit 13 and a grouping unit 14.
- the feature amount calculation means 13 calculates the feature amount of the product for each product.
- the grouping means 14 determines the customer group and the product group based on the product purchase record obtained for each combination of the customer and the product, the distribution of the product purchase record, the feature value for each product, and the feature value distribution. Determine. Specifically, the grouping means 14 obtains for each combination of a customer and a product, a purchase performance index value that is an index value indicating a product purchase performance, a distribution parameter of the purchase performance index value, and a feature amount for each product. And a group of customers and a group of products based on the distribution parameter of the feature amount.
- the grouping system 11 may define a service group.
- a grouping system comprising: a purchasing tendency calculation unit; and a grouping unit that determines a group of customers based on the trend and the distribution of the trend, and a group of products or a group of product-related items.
- a purchase tendency calculation means for every combination of a customer and goods, or for every combination of goods related matters which are matters related to a customer and goods
- the purchase trend index value which is an index value indicating the tendency of the customer to purchase the product
- the grouping means determines the customer group based on the purchase trend index value and the distribution parameter of the purchase trend index value.
- the grouping system according to appendix 1, which defines a group of products or a group of product-related items.
- the grouping means uses a plurality of sets of likelihoods obtained by combining a customer group, a product group or a product-related item group, and a distribution parameter of a purchase tendency index value,
- the grouping system according to appendix 2 which defines a group and a group of products or a group of product-related matters.
- the purchase tendency calculating means calculates price elasticity as a purchase tendency index value for each combination of a customer and a product, and the grouping means is described in Supplementary 2 or Supplementary 3 for determining a customer group and a product group Grouping system.
- the purchase trend calculation means calculates a value indicating the degree of days from the product advertisement date until the customer purchases the product as a purchase trend index value. 4.
- the purchase trend calculation means calculates a value indicating the degree of days from the product release date until the customer purchases the product as a purchase trend index value. 4.
- the purchase trend calculation means calculates a value indicating the degree of days from when the product is displayed at the store until the customer purchases the product as a purchase trend index value.
- the means is the grouping system according to the supplementary note 2 or the supplementary note 3, which defines a customer group and a product group.
- the purchase tendency calculating means calculates the average purchase interval as a purchase tendency index value for each combination of the customer and the product, and the grouping means is described in the supplementary note 2 or the supplementary note 3 that defines the customer group and the product group. Grouping system.
- the purchase trend calculation means calculates the customer's commitment to an individual product in the product category as a purchase tendency index value, and the grouping means defines a group of customers.
- the grouping system according to appendix 2 or appendix 3, wherein a group of product categories is defined as a group of product-related items.
- the purchase trend calculation means calculates the customer's attention to an individual manufacturer of the product in the product category as a purchase tendency index value
- the grouping means calculates the customer group
- the grouping system according to Supplementary Note 2 or Supplementary Note 3, wherein a group of product categories is defined as a group of product-related items.
- the purchase trend calculation means calculates the customer's attention to an individual brand of the product in the product category as a purchase tendency index value.
- the grouping means is any one of Supplementary notes 1 to 8 that defines a group of customers and a group of products so that one customer belongs to only one group and one product belongs to only one group.
- the grouping means allows one customer to belong to a plurality of groups and one product to belong to a plurality of groups, and defines a customer group and a product group.
- a grouping system according to any one of the above.
- the grouping means includes the supplementary notes 1, 2, and 7 that define a group of customers and a group of product related items such that one customer belongs to only one group and one product related item belongs to only one group.
- the grouping system according to any one of 3, 9, 10, and 11.
- the grouping means allows a customer to belong to a plurality of groups, and allows one product-related matter to belong to a plurality of groups, and defines a customer group and a product-related matter group.
- the grouping system according to any one of 1, 2, 3, 9, 10, and 11.
- the feature-value calculation means which calculates the feature-value of a product for every product, The product purchase results obtained for every combination of a customer and a product, Distribution of the said product purchase results, The said feature-value for every product,
- a grouping system comprising grouping means for determining a group of customers and a group of products based on the distribution of feature quantities.
- the grouping means includes a purchase record index value that is an index value indicating a product purchase record obtained for each combination of a customer and a product, a distribution parameter of the purchase record index value, and a feature value for each product.
- the grouping system according to attachment 16 wherein a group of customers and a group of products are defined based on the distribution parameter of the feature amount.
- the grouping means uses a plurality of sets of likelihoods obtained by combining a customer group, a product group, a distribution parameter of purchase performance index values, and a distribution parameter of the feature amount of the product.
- the grouping system according to appendix 17, which defines a customer group and a product group.
- the grouping means is any one of supplementary note 16 to supplementary note 19, which defines a customer group and a product group so that one customer belongs to only one group and one product belongs to only one group.
- the grouping means allows one customer to belong to a plurality of groups and one product to belong to a plurality of groups, and defines a customer group and a product group.
- a grouping system according to any one of the above.
- a grouping system comprising: a purchasing tendency calculation means; and a grouping means for determining a group of customers and a group of services or service-related matters based on the trend and the distribution of the trends.
- the feature-value calculation means which calculates the feature-value of a service for every service, The service purchase track record obtained for every combination of a customer and a service, Distribution of the said service purchase track record, The said feature-value for every service, A grouping system comprising grouping means for determining a group of customers and a group of services based on the distribution of feature quantities.
- the tendency for the customer to purchase the product is calculated for each combination of the customer and the product, or for each combination of the product and the product-related item that is related to the customer and the product.
- a grouping method characterized by defining a group of customers and a group of products or a group of product related items based on the trend and the distribution of the trends.
- the feature-value of a product is calculated for every product,
- the product purchase track record obtained for every combination of a customer and a product, the distribution of the said product purchase track record, the said feature-value for every product, and the distribution of the feature-value
- a grouping method characterized by defining a group of customers and a group of products based on
- the tendency for the customer to purchase the service is calculated for each combination of the customer and the service, or for each combination of the service and related items that are related to the customer and the service.
- a grouping method characterized by defining a group of customers and a group of services or a group of service-related matters based on the trend and the distribution of the trends.
- the feature-value of a service is calculated for every service, The service purchase track record obtained for every combination of a customer and a service, the distribution of the said service purchase track record, the said feature-value for every service, and the distribution of the said feature-value
- a grouping method characterized in that a group of customers and a group of services are defined based on
- a feature amount calculation process for calculating a feature amount of a product for each product, a product purchase record obtained for each combination of a customer and a product, a distribution of the product purchase record, and the product purchase result A grouping program for executing a grouping process for determining a group of customers and a group of products based on the feature amount and the distribution of the feature amount.
- the feature-value calculation process which calculates the feature-value of a service for every service in a computer, the service purchase performance obtained for every combination of a customer and a service, the distribution of the said service purchase performance, and the said for every service
- a grouping program for executing a grouping process for determining a customer group and a service group based on a feature amount and a distribution of the feature amount.
- the purchase tendency calculation means 3 is a tendency for the customer to purchase a product for each combination of the customer and the product or for each combination of the customer and the product-related matter ( Specifically, the case of calculating the purchasing tendency index value) is shown.
- the grouping system may be configured to acquire a tendency (purchasing tendency index value) of a customer to purchase a product from the outside. A configuration example of the grouping system in this case is shown in FIG.
- the grouping system 90 shown in FIG. 24 includes purchase tendency acquisition means 95 and grouping means 4.
- the purchase tendency acquisition unit 95 acquires a tendency (purchasing tendency index value) that the customer purchases the product for each combination of the customer and the product, or for each combination of the customer and the product related item. At this time, the purchasing tendency acquisition means 95 also acquires information on the type of distribution of the purchasing tendency index value. For example, on a server provided outside the grouping system 90, a purchase tendency index value determined for each combination of a customer and a product or a combination of a customer and a product-related item, and a type of distribution of the purchase trend index value Is stored in advance by the administrator. The purchase tendency acquisition unit 95 may acquire information on each purchase tendency index value and the type of distribution of the purchase tendency index value from the server via the communication network.
- the aspect in which the purchase tendency acquisition means 95 acquires a purchase tendency index value etc. from the outside is not limited to said aspect, Another aspect may be sufficient.
- the purchase tendency acquisition unit 95 may accept information on each purchase tendency index value and its distribution type input from the outside.
- the purchase trend index value acquired by the purchase trend acquisition unit 95 is the same as the purchase trend index value calculated by the purchase trend calculation unit 3 in the first embodiment.
- the grouping means 4 determines a group of customers and a group of products or a group of product related items based on the tendency of the customer to purchase products and their distribution. This operation is the same as the operation of the grouping unit 4 of the first embodiment.
- the purchase tendency acquisition unit 95 and the grouping unit 4 are realized by a CPU of a computer, for example.
- the CPU may read the grouping program from a program recording medium such as a computer program storage device (not shown in FIG. 24) and operate as the purchase tendency acquisition unit 95 and the grouping unit 4 according to the grouping program.
- a program recording medium such as a computer program storage device (not shown in FIG. 24)
- Each means may be realized by separate hardware.
- the grouping program obtains a purchase tendency for the customer to acquire a tendency to purchase the product for each combination of the customer and the product or for each combination of the product and the product-related matter that is a matter related to the customer and the product. It can be said that the program is a program for executing a grouping process for determining a group of products and a group of product-related items as well as a group of customers based on the process and the trend and the distribution of the trend.
- the purchase tendency acquisition unit 95 may acquire a tendency for a customer to purchase a service for each combination of a customer and a service, or for each combination of a customer and a service-related item.
- the grouping means 4 may determine a group of customers and a group of services or a group of service related matters based on the tendency of the customer to purchase the service and its distribution.
- the feature amount calculation unit 13 calculates the feature amount of the product for each product
- the grouping unit 14 records the product purchase record for each combination of the customer and the product ( Specifically, the case where the product purchase actual value) is calculated is shown.
- the grouping system may be configured to acquire the feature amount of the product and the product purchase record (product purchase record value) from the outside. A configuration example of the grouping system in this case is shown in FIG.
- 25 includes an information acquisition unit 96 and a grouping unit 97.
- the information acquisition unit 96 acquires the feature amount of the product for each product. At this time, the information acquisition unit 96 also acquires information on the type of distribution of the feature amount.
- the information acquisition means 96 also acquires a product purchase record value for each combination of customer and product. At this time, the information acquisition unit 96 also acquires information on the distribution type of the product purchase record value.
- the administrator stores in advance information on the feature amount of each product determined for each product and the type of distribution of the feature amount in a server provided outside the grouping system 91. Further, the manager stores in advance information on the product purchase record value determined for each combination of the customer and the product and the distribution type of the product purchase record value in the server.
- the information acquisition unit 96 receives information about the feature amount of each product and the distribution type of the feature amount, and information about the distribution type of each product purchase actual value and the product purchase actual value from the server via the communication network. And get it.
- the aspect in which the information acquisition means 96 acquires such information is not limited to the above aspect, and may be another aspect. For example, even if the information acquisition means 96 accepts information on the feature amount of each product and the type of distribution of the feature amount input from the outside, information on the type of each product purchase actual value and the distribution of the product purchase actual value, Good.
- the feature amount of the product acquired by the information acquisition unit 96 is the same as the feature amount of the product calculated by the feature amount calculation unit 13 in the second embodiment.
- the product purchase record value acquired by the information acquisition unit 96 is the same as the product purchase record value calculated by the grouping unit 14 in the second embodiment.
- the grouping means 97 determines the customer group based on the product purchase record value determined for each combination of the customer and the product, the distribution of the product purchase record value, the feature value for each product, and the feature value distribution. And define product groups. This operation is the same as the operation in which the grouping means 14 determines the customer group and the product group in the second embodiment.
- the information acquisition unit 96 and the grouping unit 97 are realized by a CPU of a computer, for example.
- the CPU may read the grouping program from a program recording medium such as a computer program storage device (not shown in FIG. 25) and operate as the information acquisition unit 96 and the grouping unit 97 according to the grouping program.
- a program recording medium such as a computer program storage device (not shown in FIG. 25)
- Each means may be realized by separate hardware.
- the grouping program causes the computer to acquire the feature amount of the product for each product, the product purchase results determined for each combination of the customer and the product, the distribution of the product purchase results, It can be said that this is a program for executing a grouping process for determining a group of customers and a group of products based on each feature quantity and the distribution of the feature quantity.
- the information acquisition means 96 includes a service feature amount determined for each service, information on the type of distribution of the feature amount, a service purchase record determined for each combination of customer and service, and a distribution of the service purchase record May be acquired.
- the grouping means 97 determines the customer based on the service purchase record obtained for each combination of the customer and the service, the distribution of the service purchase record, the feature value for each service, and the distribution of the feature value. And a group of services may be defined.
- the present invention is suitably applied to a grouping system that groups customers and items related to products or items related to services or services.
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
その傾向およびその傾向の分布に基づいて、顧客のグループを定めるとともに、サービスのグループまたはサービス関連事項のグループを定めるグルーピング手段とを備えることを特徴とする。
図1は、本発明の第1の実施形態のグルーピングシステムの例を示すブロック図である。第1の実施形態のグルーピングシステム1は、データ記憶手段2と、購買傾向算出手段3と、グルーピング手段4とを備える。
例1では、顧客が商品の値下げ(割引率でもよい。)に応じてその商品の購買量を変える度合いを、購買傾向算出手段3が購買傾向指標値として算出する。例えば、購買傾向算出手段3は、価格弾力性を購買傾向指標値として算出する。
例2では、商品の広告日から顧客が商品を購買するまでの日数の程度を示す値を、購買傾向算出手段3が購買傾向指標値として算出する。
例3では、新商品の発売日から顧客が商品を購買するまでの日数の程度を示す値を、購買傾向算出手段3が購買傾向指標値として算出する。
例4では、商品が店舗に陳列されてから顧客が商品を購買するまでの日数の程度を示す値を、購買傾向算出手段3が購買傾向指標値として算出する。
例5では、顧客がある商品を購買してから次にその商品を購買するまでの日数の程度を示す値を、購買傾向算出手段3が購買傾向指標値として算出する。
例6では、商品カテゴリ内の個別の商品への顧客のこだわり度(換言すれば、愛着度)を、購買傾向算出手段3が購買傾向指標値として算出する。
例7では、商品カテゴリ内の個別の商品のメーカへの顧客のこだわり度(換言すれば、愛着度)を、購買傾向算出手段3が購買傾向指標値として算出する。
図5は、本発明の第2の実施形態のグルーピングシステムの例を示すブロック図である。第2の実施形態のグルーピングシステム11は、データ記憶手段12と、特徴量算出手段13と、グルーピング手段14とを備える。
以下に示す具体例1では、第2の実施形態における商品の特徴量の具体例を示す。特徴量算出手段13は、商品カテゴリ毎に、カテゴリ内の各商品の特徴量を算出する。本例では、特徴量算出手段13が商品の特徴量として、相対価格を算出する場合を例にする。
具体例2では、第1の実施形態で説明した例1の具体例を示す。購買傾向算出手段3は、購買データに基づいて、顧客IDおよび商品IDの組み合わせに対して、その商品の実際の販売価格毎に、購買数(購買量)を特定する。この結果、図12に例示するように、顧客ID、商品IDおよび商品の実際の販売価格の組み合わせに対して、購買数を対応付けた情報が得られる。
具体例3では、第1の実施形態で説明した例2の具体例を示す。
具体例4では、商品の発売日からの経過日数を利用して購買傾向指標値xc,iを求める場合の例を示す。
具体例5では、第1の実施形態で説明した例5の具体例を示す。
具体例6では、第1の実施形態で説明した例6の具体例を示す。
具体例7では、第1の実施形態で説明した例7の具体例を示す。
2,12 データ記憶手段
3 購買傾向算出手段
4,14 グルーピング手段
13 特徴量算出手段
Claims (31)
- 顧客の商品購買状況に基づいて、顧客および商品の組み合わせ毎、または、顧客および、商品に関連する事項である商品関連事項の組み合わせ毎に、顧客が商品を購買する傾向を算出する購買傾向算出手段と、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、商品のグループまたは商品関連事項のグループを定めるグルーピング手段とを備える
ことを特徴とするグルーピングシステム。 - 購買傾向算出手段は、顧客の商品購買状況を示す購買データに基づいて、顧客および商品の組み合わせ毎、または、顧客および、商品に関連する事項である商品関連事項の組み合わせ毎に、顧客が商品を購買する傾向を示す指標値である購買傾向指標値を算出し、
グルーピング手段は、前記購買傾向指標値および前記購買傾向指標値の分布のパラメータに基づいて、顧客のグループを定めるとともに、商品のグループまたは商品関連事項のグループを定める
請求項1に記載のグルーピングシステム。 - グルーピング手段は、
顧客のグループと、商品のグループまたは商品関連事項のグループと、購買傾向指標値の分布のパラメータとを組み合わせて得られる複数の組の尤度を用いて、顧客のグループと、商品のグループまたは商品関連事項のグループとを定める
請求項2に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品の組み合わせ毎に価格弾力性を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループおよび商品のグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品の組み合わせ毎に、商品の広告日から顧客が商品を購買するまでの日数の程度を示す値を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループおよび商品のグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品の組み合わせ毎に、商品の発売日から顧客が商品を購買するまでの日数の程度を示す値を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループおよび商品のグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品の組み合わせ毎に、商品が店舗に陳列されてから顧客が商品を購買するまでの日数の程度を示す値を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループおよび商品のグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品の組み合わせ毎に平均購買間隔を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループおよび商品のグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品カテゴリの組み合わせ毎に、商品カテゴリ内の個別の商品への顧客のこだわり度を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループを定めるとともに、商品関連事項のグループとして商品カテゴリのグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品カテゴリの組み合わせ毎に、商品カテゴリ内の商品の個別のメーカへの顧客のこだわり度を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループを定めるとともに、商品関連事項のグループとして商品カテゴリのグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - 購買傾向算出手段は、顧客および商品カテゴリの組み合わせ毎に、商品カテゴリ内の商品の個別のブランドへの顧客のこだわり度を購買傾向指標値として算出し、
グルーピング手段は、顧客のグループを定めるとともに、商品関連事項のグループとして商品カテゴリのグループを定める
請求項2または請求項3に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が1つのグループのみに属し、1つの商品が1つのグループのみに属するように、顧客のグループおよび商品のグループを定める
請求項1から請求項8のうちのいずれか1項に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が複数のグループに属すること、および、1つの商品が複数のグループに属することを許容して、顧客のグループおよび商品のグループを定める
請求項1から請求項8のうちのいずれか1項に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が1つのグループのみに属し、1つの商品関連事項が1つのグループのみに属するように、顧客のグループおよび商品関連事項のグループを定める
請求項1、2、3、9、10、11のうちのいずれか1項に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が複数のグループに属すること、および、1つの商品関連事項が複数のグループに属することを許容して、顧客のグループおよび商品関連事項のグループを定める
請求項1、2、3、9、10、11のうちのいずれか1項に記載のグルーピングシステム。 - 商品毎に商品の特徴量を算出する特徴量算出手段と、
顧客および商品の組み合わせ毎に得られる商品購買実績と、前記商品購買実績の分布と、商品毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよび商品のグループを定めるグルーピング手段とを備える
ことを特徴とするグルーピングシステム。 - グルーピング手段は、顧客および商品の組み合わせ毎に得られる、商品購買実績を示す指標値である購買実績指標値と、前記購買実績指標値の分布のパラメータと、商品毎の特徴量と、前記特徴量の分布のパラメータとに基づいて、顧客のグループおよび商品のグループを定める
請求項16に記載のグルーピングシステム。 - グルーピング手段は、
顧客のグループと、商品のグループと、購買実績指標値の分布のパラメータと、商品の特徴量の分布のパラメータとを組み合わせて得られる複数の組の尤度を用いて、顧客のグループおよび商品のグループを定める
請求項17に記載のグルーピングシステム。 - 特徴量算出手段は、商品毎に、商品の特徴量として商品の相対価格を算出する
請求項16から請求項18のうちのいずれか1項に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が1つのグループのみに属し、1つの商品が1つのグループのみに属するように、顧客のグループおよび商品のグループを定める
請求項16から請求項19のうちのいずれか1項に記載のグルーピングシステム。 - グルーピング手段は、一人の顧客が複数のグループに属すること、および、1つの商品が複数のグループに属することを許容して、顧客のグループおよび商品のグループを定める
請求項16から請求項19のうちのいずれか1項に記載のグルーピングシステム。 - 顧客のサービス購買状況に基づいて、顧客およびサービスの組み合わせ毎、または、顧客および、サービスに関連する事項であるサービス関連事項の組み合わせ毎に、顧客がサービスを購買する傾向を算出する購買傾向算出手段と、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、サービスのグループまたはサービス関連事項のグループを定めるグルーピング手段とを備える
ことを特徴とするグルーピングシステム。 - サービス毎にサービスの特徴量を算出する特徴量算出手段と、
顧客およびサービスの組み合わせ毎に得られるサービス購買実績と、前記サービス購買実績の分布と、サービス毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよびサービスのグループを定めるグルーピング手段とを備える
ことを特徴とするグルーピングシステム。 - 顧客の商品購買状況に基づいて、顧客および商品の組み合わせ毎、または、顧客および、商品に関連する事項である商品関連事項の組み合わせ毎に、顧客が商品を購買する傾向を算出し、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、商品のグループまたは商品関連事項のグループを定める
ことを特徴とするグルーピング方法。 - 商品毎に商品の特徴量を算出し、
顧客および商品の組み合わせ毎に得られる商品購買実績と、前記商品購買実績の分布と、商品毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよび商品のグループを定める
ことを特徴とするグルーピング方法。 - 顧客のサービス購買状況に基づいて、顧客およびサービスの組み合わせ毎、または、顧客および、サービスに関連する事項であるサービス関連事項の組み合わせ毎に、顧客がサービスを購買する傾向を算出し、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、サービスのグループまたはサービス関連事項のグループを定める
ことを特徴とするグルーピング方法。 - サービス毎にサービスの特徴量を算出し、
顧客およびサービスの組み合わせ毎に得られるサービス購買実績と、前記サービス購買実績の分布と、サービス毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよびサービスのグループを定める
ことを特徴とするグルーピング方法。 - コンピュータに、
顧客の商品購買状況に基づいて、顧客および商品の組み合わせ毎、または、顧客および、商品に関連する事項である商品関連事項の組み合わせ毎に、顧客が商品を購買する傾向を算出する購買傾向算出処理、および、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、商品のグループまたは商品関連事項のグループを定めるグルーピング処理
を実行させるためのグルーピングプログラム。 - コンピュータに、
商品毎に商品の特徴量を算出する特徴量算出処理、および、
顧客および商品の組み合わせ毎に得られる商品購買実績と、前記商品購買実績の分布と、商品毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよび商品のグループを定めるグルーピング処理
を実行させるためのグルーピングプログラム。 - コンピュータに、
顧客のサービス購買状況に基づいて、顧客およびサービスの組み合わせ毎、または、顧客および、サービスに関連する事項であるサービス関連事項の組み合わせ毎に、顧客がサービスを購買する傾向を算出する購買傾向算出処理、および、
前記傾向および前記傾向の分布に基づいて、顧客のグループを定めるとともに、サービスのグループまたはサービス関連事項のグループを定めるグルーピング処理
を実行させるためのグルーピングプログラム。 - コンピュータに、
サービス毎にサービスの特徴量を算出する特徴量算出処理、および、
顧客およびサービスの組み合わせ毎に得られるサービス購買実績と、前記サービス購買実績の分布と、サービス毎の前記特徴量と、前記特徴量の分布とに基づいて、顧客のグループおよびサービスのグループを定めるグルーピング処理
を実行させるためのグルーピングプログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US15/534,273 US20180260829A1 (en) | 2014-12-10 | 2015-11-27 | Grouping system, grouping method, and grouping program |
JP2016563402A JP6662298B2 (ja) | 2014-12-10 | 2015-11-27 | グルーピングシステム、グルーピング方法およびグルーピングプログラム |
EP15866581.0A EP3232390A4 (en) | 2014-12-10 | 2015-11-27 | Grouping system, grouping method, and grouping program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2014249953 | 2014-12-10 | ||
JP2014-249953 | 2014-12-10 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016092767A1 true WO2016092767A1 (ja) | 2016-06-16 |
Family
ID=56106995
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2015/005926 WO2016092767A1 (ja) | 2014-12-10 | 2015-11-27 | グルーピングシステム、グルーピング方法およびグルーピングプログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US20180260829A1 (ja) |
EP (1) | EP3232390A4 (ja) |
JP (1) | JP6662298B2 (ja) |
WO (1) | WO2016092767A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11188568B2 (en) | 2016-11-14 | 2021-11-30 | Nec Corporation | Prediction model generation system, method, and program |
US11301763B2 (en) | 2016-11-14 | 2022-04-12 | Nec Corporation | Prediction model generation system, method, and program |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001331628A (ja) * | 2000-05-24 | 2001-11-30 | Just Syst Corp | マーケティング調査システム及び方法、装置並びに記録媒体 |
JP2002073946A (ja) * | 2000-08-29 | 2002-03-12 | Ikeda Keiei Design Kenkyusho:Kk | 販売促進支援方法および販売促進支援装置 |
JP2007272571A (ja) * | 2006-03-31 | 2007-10-18 | Nomura Research Institute Ltd | マーケティングリサーチ支援サーバ及び方法 |
JP2008516355A (ja) * | 2004-10-13 | 2008-05-15 | ダンハンビー リミテッド | 小売店における製品の価格を決定するための方法 |
JP2008186413A (ja) * | 2007-01-31 | 2008-08-14 | Ntt Data Corp | 需要予測装置、需要予測方法、及び、需要予測プログラム |
JP2009134487A (ja) * | 2007-11-30 | 2009-06-18 | Hitachi Software Eng Co Ltd | 消費財購入情報提供システム |
JP2014078179A (ja) * | 2012-10-11 | 2014-05-01 | Dentsu Inc | 情報処理装置、情報処理プログラム及び情報処理方法 |
-
2015
- 2015-11-27 JP JP2016563402A patent/JP6662298B2/ja active Active
- 2015-11-27 US US15/534,273 patent/US20180260829A1/en not_active Abandoned
- 2015-11-27 WO PCT/JP2015/005926 patent/WO2016092767A1/ja active Application Filing
- 2015-11-27 EP EP15866581.0A patent/EP3232390A4/en not_active Withdrawn
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001331628A (ja) * | 2000-05-24 | 2001-11-30 | Just Syst Corp | マーケティング調査システム及び方法、装置並びに記録媒体 |
JP2002073946A (ja) * | 2000-08-29 | 2002-03-12 | Ikeda Keiei Design Kenkyusho:Kk | 販売促進支援方法および販売促進支援装置 |
JP2008516355A (ja) * | 2004-10-13 | 2008-05-15 | ダンハンビー リミテッド | 小売店における製品の価格を決定するための方法 |
JP2007272571A (ja) * | 2006-03-31 | 2007-10-18 | Nomura Research Institute Ltd | マーケティングリサーチ支援サーバ及び方法 |
JP2008186413A (ja) * | 2007-01-31 | 2008-08-14 | Ntt Data Corp | 需要予測装置、需要予測方法、及び、需要予測プログラム |
JP2009134487A (ja) * | 2007-11-30 | 2009-06-18 | Hitachi Software Eng Co Ltd | 消費財購入情報提供システム |
JP2014078179A (ja) * | 2012-10-11 | 2014-05-01 | Dentsu Inc | 情報処理装置、情報処理プログラム及び情報処理方法 |
Non-Patent Citations (1)
Title |
---|
See also references of EP3232390A4 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11188568B2 (en) | 2016-11-14 | 2021-11-30 | Nec Corporation | Prediction model generation system, method, and program |
US11301763B2 (en) | 2016-11-14 | 2022-04-12 | Nec Corporation | Prediction model generation system, method, and program |
Also Published As
Publication number | Publication date |
---|---|
JPWO2016092767A1 (ja) | 2017-10-05 |
EP3232390A4 (en) | 2018-04-18 |
EP3232390A1 (en) | 2017-10-18 |
US20180260829A1 (en) | 2018-09-13 |
JP6662298B2 (ja) | 2020-03-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP6424225B2 (ja) | 小売施設内で消費者の位置を予測するシステム及び方法 | |
Bustos-Reyes et al. | Store and store format loyalty measures based on budget allocation | |
JP6352798B2 (ja) | マーケティング施策最適化装置、方法、及びプログラム | |
US20160189177A1 (en) | Determination of a Purchase Recommendation | |
JP2009205365A (ja) | 商品の在庫管理および販売の最適化システム、その最適化方法、及びその最適化プログラム | |
Casado et al. | Consumer price sensitivity in the retail industry: latitude of acceptance with heterogeneous demand | |
Lee et al. | Consumers' choice for fresh food at online shopping in the time of covid19 | |
Bockholdt et al. | Private label shoppers between fast fashion trends and status symbolism–A customer characteristics investigation | |
Herstein et al. | Private and national brand consumers' images of fashion stores | |
Stan | Antecedents of customer loyalty in the retailing sector: the impact of switching costs | |
Deka | Factors Influencing Consumers' Choice of Retail Store Format in Assam, India. | |
Hillen | Psychological pricing in online food retail | |
Jung et al. | A meta-analysis of correlations between market share and other brand performance metrics in FMCG markets | |
Sebri et al. | Estimating umbrella-branding spillovers: a retailer perspective | |
WO2016092767A1 (ja) | グルーピングシステム、グルーピング方法およびグルーピングプログラム | |
JP6699652B2 (ja) | グルーピングシステム、方法、およびプログラム | |
Smith et al. | Predictive analytics improves sales forecasts for a pop-up retailer | |
JP2016081199A (ja) | 広告配信システム | |
JP6287280B2 (ja) | 情報処理方法、プログラム、及び情報処理装置 | |
Atulkar et al. | Adoption of retailer centric philosophy in organizing buying process: a case study on Indian buying systems | |
Chowdhury et al. | Antecedents and consequences of customer satisfaction: An empirical study on retail store in Bangladesh | |
Zhuang et al. | Consumer Choice of Private Label or National Brand: The case of organic and non-organic milk | |
Verhelst et al. | Implicit contracts and price stickiness: evidence from customer-level scanner data | |
Jacob et al. | The Determinant Factors that Influence Repurchase Intention of Samsung Smartphone in Jabodetabek | |
US11948181B2 (en) | Systems and methods for using SKU vector information |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15866581 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2016563402 Country of ref document: JP Kind code of ref document: A |
|
REEP | Request for entry into the european phase |
Ref document number: 2015866581 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 15534273 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |