JP5155290B2 - Purchase stage determination apparatus and purchase stage determination method - Google Patents

Description

  The present invention relates to a purchase stage determination apparatus and a purchase stage determination method for determining whether a user state is a purchase stage.

  Conventionally, a technique for distributing advertisement information to a user's terminal device based on the purchase history of the user has been proposed (see, for example, Patent Document 1).

  Patent Document 1 proposes a technique for distributing an advertisement at an appropriate time for each user using the purchase history of the user such as product information, personal information, and regional information.

JP 2004-206415 A

  However, in the technique described in Patent Literature 1, if there is no purchase history by the user, whether or not the user is highly likely to perform the purchase of goods or the use of services by the user, that is, the user There was a problem that it was not possible to determine whether the situation was in the purchasing stage.

  Accordingly, an object of the present invention is to provide a purchase stage determination device and a purchase stage determination method for determining whether a user's situation is a purchase stage even when a purchase history is not used.

  (1) A purchase stage determination device for determining whether the state of the user is a purchase stage based on a query used for a search by a terminal device corresponding to the user, the identification information for identifying the user, and the A search history acquisition means for acquiring a search history including a query, a reception means for receiving a category and a keyword belonging to the category as teacher data, and a learning model based on the category and the keyword received by the reception means. And generating a keyword based on the query by classifying the query included in the search history by a classifier using the generated learning model and using the classified query as a new keyword of the category. A keyword generating means; and the category received by the receiving means; A storage unit that associates and stores the keyword generated by the keyword generation unit, and a query that is continuously used within a predetermined period for a search by the terminal device corresponding to the user identified by the identification information The query set extraction means for extracting one or more from the search history as a unit, and the query extracted by the query set extraction means based on the category and the keyword stored in the storage means Based on the category classification means for classifying a set for each category, the set of queries extracted by the query set extraction means, and the category in which the set of queries is classified by the category classification means, for each category A use frequency calculating means for calculating a use frequency, and the use frequency calculating means A graph generation unit that generates a graph of a predetermined format using the frequency of use of the category that has been issued, and determines whether the graph generated by the graph generation unit is in a predetermined state in which a certain category can be identified And a determination unit that determines, when the graph is in the predetermined state, that the state of the user is the purchase stage of the category corresponding to the predetermined state.

  In the purchase stage determination apparatus of (1), the search history acquisition unit acquires a search history including identification information for identifying the user and a query used for the search by the user, and the reception unit receives the category and category as teacher data. Accept keywords belonging to. The keyword generation unit generates a learning model based on the category and the keyword received by the reception unit, classifies the query included in the search history by a classifier using the generated learning model, and newly sets the classified query. A keyword based on the query is generated as a unique keyword. The storage means stores the category received by the receiving means in association with the keyword generated by the keyword generating means, and the query set extracting means performs a search by the terminal device corresponding to the user identified by the identification information. One or a plurality of queries are extracted from the search history with a set of queries used continuously within a predetermined period as a unit. The category classification means classifies the set of queries extracted by the query set extraction means for each category based on the categories and keywords stored in the storage means, and the usage frequency calculation means is extracted by the query set extraction means. The frequency of use for each category is calculated based on the set of queries and the category in which the set of queries is classified by the category classification means. The graph generation means generates a graph in a predetermined format using the category usage frequency calculated by the usage frequency calculation means, and the determination means determines whether the graph generated by the graph generation means is in a predetermined state. If the graph is in a predetermined state, it is determined that the user state is the category purchase stage.

  Thereby, the purchase stage determination device determines whether a state having a higher use frequency than the use frequency of other categories appears a predetermined number of times in the graph generated by the graph generation unit, for example. It is possible to determine whether or not the user is highly likely to perform the purchase of the product or the use of the service, that is, whether or not the user is in the purchase stage.

  (2) An advertisement information storage unit that stores advertisement information in association with the category; and when the determination unit determines that the state of the predetermined user is the purchase stage, from the advertisement information storage unit, The purchase stage determination according to (1), further comprising: advertisement distribution means for extracting the advertisement information corresponding to the category used for generation and distributing the extracted advertisement information to the terminal device corresponding to the user. apparatus.

  In the purchase stage determination apparatus of (2), the advertisement distribution means generates a graph from the advertisement information storage means when the determination means determines that the graph is in a predetermined state and the predetermined user state is the purchase stage. The advertisement information corresponding to the category used in the above is extracted, and the extracted advertisement information is distributed to the terminal device corresponding to the user.

  Thereby, the purchase stage determination apparatus can distribute advertisement information corresponding to this category to a terminal apparatus corresponding to a user who is highly likely to purchase a product related to a certain category or use a service.

  (3) A purchase stage determination method for determining whether the state of the user is a purchase stage based on a query used for a search by a terminal device corresponding to the user, the identification information identifying the user, and the A search history acquisition step of acquiring a search history including a query, a reception step of receiving a category and a keyword belonging to the category as teacher data, and a learning model based on the category and the keyword received in the reception step And generating a keyword based on the query by classifying the query included in the search history by a classifier using the generated learning model and using the classified query as a new keyword of the category. Received in the keyword generation step and the reception step In a predetermined period of time, the storage step of storing the associated category and the keyword generated in the keyword generation step in association with each other and the search by the terminal device corresponding to the user identified by the identification information A query set extraction step for extracting one or a plurality of queries from the search history with a set of queries continuously used as a unit, and the query set extraction step based on the category and the keyword stored in the storage step A category classification step for classifying the set of queries extracted in each category, the set of queries extracted in the query set extraction step and the set of queries into the categories classified in the category classification step Based on the above categories A usage frequency calculation step for calculating a usage frequency of the category, a graph generation step for generating a graph in a predetermined format using the usage frequency for each category calculated in the usage frequency calculation step, and a graph generation step It is determined whether or not the graph is in a predetermined state in which a certain category can be specified, and when the graph is in the predetermined state, the purchase state of the category corresponding to the predetermined state is the user state. A purchase stage determination method for causing a computer to execute a determination step of determining that a stage is present.

  According to the purchase stage determination method of (3), an invention having the same operational effects as the purchase stage determination apparatus of (1) can be provided as a method.

  ADVANTAGE OF THE INVENTION According to this invention, even if it is a case where a purchase history is not used, the purchase stage determination apparatus and the purchase stage determination method which can determine a user's purchase action can be provided.

It is a schematic diagram which shows the whole structure of the purchase stage determination apparatus 1. FIG. It is a figure which shows an example of the search history stored in search history DB21. It is a figure which shows the teacher data table 111a as a specific example of teacher data. It is a figure which shows the keyword table 111b which is a specific example of the query memorize | stored in category DB111. It is a figure which shows the usage frequency for every category to which the collection of queries belongs. It is a graph which shows the use frequency of the query for every category. It is a flowchart shown about the process which produces | generates a keyword by the purchase stage determination apparatus 1. FIG. It is a flowchart shown about the process which determines a purchase stage by the purchase stage determination apparatus 1, and distributes an advertisement.

[overall structure]
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic diagram showing an overall configuration of a purchase stage determination apparatus 1 according to an embodiment of the present invention. The purchase stage determination apparatus 1 according to the present embodiment includes a control unit 10 and a storage unit 11.

  The control unit 10 includes a CPU and the like, and controls the entire purchase stage determination apparatus 1. The control unit 10 includes a search history acquisition unit 101, a teacher data reception unit 102, a keyword generation unit 103, a query set extraction unit 104, a category classification unit 105, a usage frequency calculation unit 106, a graph generation unit 107, The determination unit 108 and the advertisement distribution unit 109 are provided.

  The storage unit 11 includes a storage device such as a memory and a hard disk, and stores various information and various programs executed by the purchase stage determination device 1. The storage unit 11 includes a category DB 111 and an advertisement DB 112.

  The search history acquisition unit 101 acquires from the search server 2 a search history including a user identification ID as identification information for identifying the user and a query used for the search by the terminal device 3 corresponding to the user. Specifically, the search history acquisition unit 101 acquires a search history stored in the search history DB 21 of the search server 2. The information amount of the search history acquired by the search history acquisition unit 101 is determined as appropriate. Moreover, the terminal device 3 corresponding to the user refers to, for example, a terminal device identified by a user identification ID that identifies the user. Further, the search history acquisition unit 101 may acquire the search history from the cookie corresponding to the browser of the terminal device 3 corresponding to the user, instead of the search history DB 21. That is, the user identification ID and the query used by the terminal device 3 corresponding to the user are stored in the browser cookie, and the search history acquisition unit 101 uses the user identification ID and query included in the browser cookie as the search history. You may get it.

  FIG. 2 is a diagram illustrating an example of a search history stored in the search history DB 21. As shown in FIG. 2, the search history DB 21 stores a user identification ID, a session ID, and a query in association with each other as a search history. The search history is not limited to this, and may be included in, for example, a session log, a click log, or the like.

  The teacher data receiving unit 102 receives, as teacher data, a category and a keyword belonging to this category from the search server 2, for example. FIG. 3 is a diagram showing a teacher data table 111a as a specific example of teacher data. Specifically, the teacher data receiving unit 102 has categories as shown in FIG. 3 (for example, six categories of car, digital, animation, game, person, and local) and no category. Further, the teacher data receiving unit 102 receives keywords belonging to this category (for example, “Perius, Gaudi, Ponda” in the category “car”).

  The keyword generation unit 103 generates a learning model based on the category and keyword received by the teacher data reception unit 102. Then, the keyword generation unit 103 classifies the query included in the search history into a category by the classifier using the generated learning model, and sets the classified query as a new keyword of this category, thereby generating a keyword based on the query. For each category. For example, the keyword may be a query included in the search history.

  Specifically, the keyword generation unit 103 causes the search engine to execute a search using each keyword belonging to each category. Then, the keyword generation unit 103 acquires titles and summaries (snippets) having higher display ranks (for example, 200 items from 1 to 200) in the search results obtained by the search being executed by the search engine. . Here, the snippet is obtained by extracting text information including the query (keyword) used for the search in the web page of the search result. Note that the above-described classifier and search engine may be included in the purchase stage determination device 1 or may be included in an external server.

  The keyword generation unit 103 causes the classifier to machine-learn the acquired title and summary to generate a learning model. The keyword generation unit 103 classifies one or more queries included in the search history acquired by the search history acquisition unit 101 with a classifier to which the learning model is applied, and selects keywords based on the queries included in the search history. Generate. Then, the keyword generation unit 103 associates the category into which the query included in the search history is classified and the generated keyword and stores them in the category DB 111. Note that when there are a plurality of queries included in the search history, the keywords corresponding to the number of queries are generated. That is, when the query included in the search history is a keyword, the query included in the search history is composed of one or a plurality of keywords. The keyword generation unit 103 may use a classification device that performs classification using data such as a dictionary instead of the classifier.

  FIG. 4 is a diagram showing a keyword table 111b which is a specific example of categories and keywords stored in the category DB 111. As shown in FIG. In FIG. 4, keywords generated by the keyword generation unit 103 are included as compared to FIG. 3. For example, in FIG. 3, the keywords “Perius”, “Gaudi”, and “Ponda” for the category “car” are stored. On the other hand, in FIG. 4, in addition to the keywords “Perius”, “Gaudi” and “Ponda” for the category “car”, keywords such as “eco-car” and “used car” are added. Yes.

  The query set extraction unit 104 continues from the search history acquired by the search history acquisition unit 101 within a predetermined period by the terminal device 3a (see FIG. 1) corresponding to the predetermined user identified by the user identification ID. As a unit, a set of queries used as a unit is extracted. For example, the query set extraction unit 104 uses the search history acquired by the search history acquisition unit 101 to continuously use a query with a user identification ID “AAA” within a predetermined period (for example, one month). As a unit, one or more are extracted.

The query set extraction unit 104 executes the following processing for each set of queries when extracting one or a plurality of sets of queries as a unit within a predetermined period (for example, one month). It will be realized.
The query set extraction unit 104 measures the time when the query is input by the user having the user identification ID “AAA” for each query, and inputs the query (for example, query X) and the query (for example, query Y). When the interval is within time T (for example, 5 minutes), the query Y is extracted as a query included in the set of queries to which the query X belongs. The time T is not limited to 5 minutes and can be set as appropriate.

  As a more specific example, query A is input at 15:00, query B is input at 15:05, query C is input at 15:08, query D is input at 15:08, A case where the query E is input at 15:15 and the query F is input at 15:17 will be described.

  In this case, the query set extraction unit 104 sets the queries A, B, C, and D whose query input intervals are within 5 minutes as queries included in the query set G1. Moreover, the query set extraction unit 104 sets the query E as a query included in the query set G2 because the input interval between the query D and the query E is 5 minutes or more. Furthermore, since the input interval between the query E and the query F is within 5 minutes, the query set extraction unit 104 sets the query F as a query included in the query set G2.

  The category classification unit 105 classifies the query set extracted by the query set extraction unit 104 for each category based on the categories and keywords stored in the category DB 111. Specifically, the category classification unit 105 determines whether a keyword that matches a query included in the query set extracted by the query set extraction unit 104 exists in the category DB 111.

  Subsequently, if there is a keyword that matches a part or all of the queries included in the query set extracted by the query set extraction unit 104, the category classification unit 105 sets the query set to the query set. Classify into categories to which keywords that match queries included in the set belong. Here, when there are a plurality of keywords that match a query included in the query set, the category classification unit 105 sets the query set to the category to which the keyword with a high frequency belongs based on the keyword frequency. Classify. Further, when there is no keyword that matches the query set, the category classification unit 105 classifies the category of the query set as “no category”.

  The usage frequency calculation unit 106 calculates the usage frequency for each category to which the query set belongs, based on the query set extracted by the query set extraction unit 104 and the category in which the query set is classified by the category classification unit 105. To do. Specifically, in each category, the usage frequency calculation unit 106 classifies the query set extracted by the query set extraction unit 104 into this category by the category classification unit 105 within a predetermined period (for example, one day). The obtained number is calculated as the usage frequency for each category to which the set of queries belongs.

  FIG. 5 is a diagram showing the frequency of use for each category to which a set of queries belongs. As shown in FIG. 5, the usage frequency of categories such as “car”, “digital”, and “animation” is calculated for each date as the usage frequency of the category to which the set of queries by the user with the user identification ID “AAA” belongs. Has been. For example, the usage frequency of the category “car” on May 30, 2009 is “6”. That is, on May 30, 2009, there are six sets of queries belonging to the category “car” input by the user with the user identification ID “AAA”. Further, on May 30, 2009, a search with a user identification ID “AAA” was continuously performed 12 times within a predetermined period.

  The graph generation unit 107 generates a graph in a predetermined format using the category usage frequency calculated by the usage frequency calculation unit 106. Specifically, the graph generation unit 107 generates a line graph as shown in FIG. 6 using the usage frequency of the category to which the query set calculated by the usage frequency calculation unit 106 belongs. In the graphs shown in FIGS. 6A and 6B, the vertical axis indicates the usage frequency, and the horizontal axis indicates the date.

  The determination unit 108 determines whether the graph generated by the graph generation unit 107 is in a predetermined state. Specifically, in the graph generated by the graph generation unit 107, the determination unit 108 determines that a state in which the use frequency is higher than the use frequency of other categories within a predetermined period (for example, one week) a predetermined number of times ( For example, it is determined whether or not it appears.

  Specifically, as shown in FIG. 6 (a), the determination unit 108 determines that the category “car (in FIG. 6) during the period from June 8, 2009 to June 14, 2009 (one week). The state where the usage frequency of “(1)” is higher than the usage frequency of any of the other categories (♦ and Δ in FIG. 6) appears once (June 9, 2009). In this case, the determination unit 108 determines that the graph generated by the graph generation unit 107 is not in a predetermined state.

  On the other hand, as shown in FIG. 6B, the determination unit 108 determines that the frequency of use of the category “car” is another category between June 8, 2009 and June 14, 2009 (one week). 3 times (June 9, 2009, June 11, 2009, and June 12, 2009) have appeared. In this case, the determination unit 108 determines that the graph generated by the graph generation unit 107 is in a predetermined state.

  Here, when the usage frequency of the predetermined category is higher than the usage frequency of any day of other categories only once, the user's interest or interest in the predetermined category suddenly increases, etc. Can be considered.

  On the other hand, when a state in which the use frequency of a predetermined category is higher than the use frequency of any other category appears more than once within a predetermined period, the user's interest and interest in the predetermined category is time-series. It is thought that it is continuing at. The fact that the user's interest and interest in the predetermined category continues in time series is considered to be a state in which the user is highly likely to perform purchases of products and use of services related to the predetermined category.

  Therefore, the determination unit 108 determines whether or not a state in which the use frequency is higher than the use frequency of any day in other categories appears within a predetermined period in the graph generated by the graph generation unit 107. By doing so, whether or not the user is highly likely to purchase a product or use a service related to a predetermined category, that is, whether or not the user's state is a purchase stage that is causing purchase behavior Can be determined.

  When the determination unit 108 determines that the graph is in a predetermined state, the advertisement distribution unit 109 extracts advertisement information corresponding to a specific category in which the user state is determined to be in the purchase stage from the advertisement DB 112. The extracted advertisement information is distributed to the terminal device 3 (see FIG. 1) corresponding to a predetermined user. In this embodiment, the advertisement distribution unit 109 distributes the extracted advertisement information to the terminal device 3 corresponding to the user. However, the advertisement distribution unit 109 is not limited to this, and may distribute it to other servers and terminal devices.

  In the advertisement DB 112, advertisement information is stored for each category. As the advertisement information, for example, various kinds of advertisement information such as a banner advertisement, a text advertisement, an email advertisement, a moving image (streaming) advertisement, and a pop-up advertisement can be used.

  Next, processing by the purchase stage determination apparatus 1 of the present embodiment will be described. FIG. 7 is a flowchart showing a process for generating a keyword by the purchase stage determination device 1 of the present embodiment.

In step S1, the teacher data receiving unit 102 receives a category and keywords belonging to this category as teacher data.
In step S <b> 2, the search history acquisition unit 101 acquires, from the search server 2, a search history including a user identification ID and a query used for a search by the terminal device 3 corresponding to the user identification ID.

In step S3, the keyword generation unit 103 causes the search engine to execute a search using each keyword belonging to each category.
In step S4, the keyword generation unit 103 displays titles and summaries (snippets) with the highest display order (for example, 200 items from 1 to 200) in the search result obtained by executing the search by the search engine. get.

In step S5, the keyword generation unit 103 causes the classifier to machine-learn the acquired title and summary to generate a learning model.
In step S6, the keyword generation unit 103 classifies the query included in the search history acquired by the search history acquisition unit 101 into a category by a classifier to which the learning model is applied, and classifies the query as a new keyword of the category. Thus, a new keyword based on the query is generated.

In step S <b> 7, the keyword generation unit 103 stores the received category and the generated keyword in the category DB 111 in association with each other.
In this way, the purchase stage determination apparatus 1 classifies the queries included in the search history acquired by the search history acquisition unit 101 into categories by the classifier to which the learning model is applied, and converts the classified queries into new categories. By creating a keyword, a keyword based on the query is generated for each category, and the accepted category and the generated keyword are stored in association with each other.

FIG. 8 is a flowchart showing a process for determining that the user is in the purchase stage by the purchase stage determination apparatus 1 of the present embodiment and distributing the advertisement.
In step S <b> 11, the search history acquisition unit 101 acquires, from the search server 2, a search history including a user identification ID as identification information for identifying the user and a query used for the search by the terminal device 3 corresponding to the user. Note that the search history acquired in step S11 is different from the search history acquired in step S2.

In step S12, the query set extraction unit 104 extracts a set of queries used by the terminal device 3a of a predetermined user from the search history acquired in step S11.
In step S <b> 13, the category classification unit 105 determines whether a keyword that matches a query included in the query set extracted by the query set extraction unit 104 exists in the category DB 111. If it is determined that there is a keyword that matches the query included in the set of queries, the process proceeds to step S14. On the other hand, if it is determined that there is no keyword matching the query included in the query set, the category classification unit 105 proceeds to step S15.

In step S14, the category classification unit 105 classifies the query set extracted by the query set extraction unit 104 into categories to which keywords that match the queries included in the query set belong.
In step S15, the category classification unit 105 classifies the category of the query set extracted by the query set extraction unit 104 as “no category”.

In step S <b> 16, the usage frequency calculation unit 106 calculates the usage frequency for each category based on the query set extracted by the query set extraction unit 104 and the category in which the query set is classified by the category classification unit 105. .
In step S <b> 17, the graph generation unit 107 generates a graph in a predetermined format using the usage frequency for each category calculated by the usage frequency calculation unit 106.

  In step S18, the determination unit 108 determines whether or not a predetermined period of the graph generated by the graph generation unit 107 is in a predetermined state. If it is determined that the predetermined period of the graph is in the predetermined state, that is, if it is determined that the state of the predetermined user is the purchase stage, the process proceeds to step S19. On the other hand, if it is determined that the graph is not in a predetermined state, the process ends.

  In step S19, when the determination unit 108 determines that the graph is in a predetermined state, the advertisement distribution unit 109 extracts advertisement information corresponding to the category used for generating the graph from the advertisement DB 112, The extracted advertisement information is distributed to the terminal device 3 (see FIG. 1) corresponding to a predetermined user.

According to the purchase stage determination apparatus 1 of the present embodiment, the following effects can be achieved.
In the purchase stage determination device 1, the query set extraction unit 104 extracts a set of queries used for search by the terminal device 3 a corresponding to a predetermined user from the search history, and the category classification unit 105 is stored in the category DB 111. Based on the categories and keywords, the query set extracted by the query set extraction unit 104 is classified for each category. The usage frequency calculation unit 106 calculates the usage frequency for each category based on the set of extracted queries and the categories in which the query set is classified by the category classification unit 105, and the graph generation unit 107 calculates A graph of a predetermined format is generated using the usage frequency. The determination unit 108 determines whether or not a predetermined period of the generated graph is in a predetermined state, and if the predetermined period of the generated graph is in a predetermined state, the state of the predetermined user is specified It is determined that the purchase stage of the category.

  Thereby, as described above, the purchase stage determination apparatus 1 uses the predetermined category at a higher frequency than the use frequency of any other category within the predetermined period in the graph generated by the graph generation unit 107. By determining whether or not a state appears a predetermined number of times, whether or not it is highly likely that the user will purchase a product or use a service related to a predetermined category. It can be determined whether or not the purchase stage is taking action.

  In the purchase stage determination apparatus 1, the advertisement distribution unit 109 generates a graph from the advertisement DB 112 when the determination unit 108 determines that the graph is in a predetermined state and the predetermined user state is the purchase stage. Advertisement information corresponding to the category used is extracted, and the extracted advertisement information is distributed to the terminal device 3a corresponding to a predetermined user.

  Thereby, the purchase stage determination apparatus 1 can distribute advertisement information to a user who is highly likely to purchase a product related to a predetermined category or use a service.

  In addition, the computer as used in this embodiment means the information processing apparatus provided with the memory | storage part, control part, etc. which were mentioned above, the purchase stage determination apparatus 1, the search server 2, and the terminal device 3 (3a) are a memory | storage part, control. An information processing apparatus including a unit is configured.

  In the above-described embodiment, the purchase stage determination device 1, the search server 2, and the terminal device 3 have been mainly described. However, the above description has been given by installing a program in a computer and operating the computer as a terminal or a server. Functions can also be realized. Therefore, the functions realized by the purchase stage determination device 1, the search server 2, and the terminal device 3 described as an embodiment in the present invention are executed by executing the above-described method by the computer, or by executing the above-described program on the computer. It can also be realized by introducing and executing the above.

  Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and can be modified as appropriate. For example, in the above-described embodiment, the determination unit 108 determines using the graph generated by the graph generation unit 107, but the present invention is not limited to this. For example, instead of the graph generated by the graph generation unit 107, the determination unit 108 determines whether or not the use frequency of the query set for each category exceeds a threshold value a predetermined number of times (for example, two times) or more within a predetermined period. You may judge.

  In the above-described embodiment, the determination unit 108 uses whether or not the predetermined period of the graph is in a predetermined state, but the present invention is not limited to this. For example, when the query set extraction unit 104 extracts a query set from the search history, the query set may be extracted from the search history within the predetermined period described above. Further, when the graph generation unit 107 generates a graph, a graph regarding the frequency of use of the query in the predetermined period described above may be generated.

  In the embodiment described above, the sales stage determination device 1 distributes the advertisement information to the terminal device 3 corresponding to the user determined to be the purchase stage by the advertisement distribution unit 109, but the present invention is not limited thereto. . For example, in the purchase stage determination apparatus 1, the control unit 10 is included in a set of queries at the usage frequency that has been determined to be the purchase stage when the user state is determined to be the purchase stage. When a search is executed using a query in the same category as the query, keywords suitable for the query (for example, “price”, “sales”, etc.) are added and displayed on the terminal device 3 corresponding to the user.

  And the control part 10 may perform the search by a search engine using a query and the keyword added to the query, and may display the result of the performed search. Thereby, when it is determined that the user's state is the purchase stage, the purchase stage determination device 1 not only distributes the advertisement information but also displays a general search result on the terminal device 3 corresponding to the user. In this case, a more suitable search result can be displayed.

  The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

DESCRIPTION OF SYMBOLS 1 Purchase stage determination apparatus 101 Search history acquisition part 102 Teacher data reception part 103 Keyword generation part 104 Query set extraction part 105 Category classification | category part 106 Usage frequency calculation part 107 Graph generation part 108 Judgment part 109 Advertisement distribution part 111 Category DB
112 Advertising DB

Claims (3)

A purchase stage determination device that determines whether the state of the user is a purchase stage based on a query used for a search by a terminal device corresponding to the user,
Search history acquisition means for acquiring identification information for identifying the user and a search history including the query;
Accepting means for receiving a category and a keyword belonging to the category as teacher data;
Based on the category and the keyword received by the receiving means, a learning model is generated, the query included in the search history is classified by a classifier using the generated learning model, and the classified query A keyword generating means for generating a keyword based on the query for each category by setting as a new keyword of the category ;
Storage means for storing the category received by the receiving means in association with the keyword generated by the keyword generating means;
The search by the terminal device corresponding to the user identified by the identification information, said set of queries used consecutively in a predetermined time period as a unit, extracting query set to one or more extracted from the search history Means,
Category classification means for classifying the set of queries extracted by the query set extraction means for each category based on the category and the keyword stored in the storage means;
A usage frequency calculation unit that calculates a usage frequency for each category based on the set of queries extracted by the query set extraction unit and the category in which the set of queries is classified by the category classification unit;
Graph generating means for generating a graph of a predetermined format using the usage frequency for each category calculated by the usage frequency calculating means;
In the graph generated by the graph generating unit relates a category that frequently used state occurs more than a predetermined number of times than other categories within the prescribed period, the state of the user is in front Symbol purchasing stage A purchase stage determination device comprising determination means for determination. Advertisement information storage means for storing advertisement information in association with the category;
When the determination unit determines that the state of the predetermined user is the purchase stage, the advertisement information corresponding to the category used to generate the graph is extracted from the advertisement information storage unit and extracted. The purchase stage determination device according to claim 1, further comprising: an advertisement distribution unit that distributes the advertisement information to the terminal device corresponding to the user. A purchase stage determination method for determining whether the state of the user is a purchase stage based on a query used for a search by a terminal device corresponding to the user,
A search history acquisition step for acquiring identification information for identifying the user and a search history including the query;
A reception step of receiving a category and a keyword belonging to the category as teacher data;
Based on the category and the keyword received in the receiving step, a learning model is generated, the query included in the search history is classified by a classifier using the generated learning model, and the classified query A keyword generation step of generating a keyword based on the query for each category by setting as a new keyword of the category ;
A storage step of storing the category received in the reception step in association with the keyword generated in the keyword generation step;
The search by the terminal device corresponding to the user identified by the identification information, said set of queries used consecutively in a predetermined time period as a unit, extracting query set to one or more extracted from the search history Steps,
A category classification step for classifying the set of queries extracted in the query set extraction step for each category based on the category and the keyword stored in the storage step;
A usage frequency calculation step of calculating a usage frequency for each category based on the set of queries extracted in the query set extraction step and the category in which the set of queries is classified in the category classification step;
A graph generation step of generating a graph in a predetermined format using the use frequency for each category calculated in the use frequency calculation step;
In the graph generated in the graph generation step relates a category that frequently used state occurs more than a predetermined number of times than other categories within the prescribed period, the state of the user is in front Symbol purchasing stage A purchase stage determination method for causing a computer to execute a determination step for determination.

Images