JP6900158B2 - Classification device, classification method and classification program - Google Patents

Description

The present invention relates to a classification device, a classification method and a classification program.

In recent years, with the rapid spread of the Internet, information distribution via the Internet has been actively carried out. As an example of such information distribution, there is known a technique of distributing various guidance information for guiding a user to a predetermined content such as an advertisement or for guiding to a conversion such as purchasing a product. For example, there is known a technique of classifying users into a plurality of groups according to attribute information and distributing guidance information to users classified into a predetermined group.

Japanese Unexamined Patent Publication No. 2008-299542 Patent No. 4808207

However, in the above-mentioned conventional technology, there are cases where users cannot be properly classified.

For example, in the above-mentioned conventional technique, each user is only classified according to the correspondence between the preset attribute information and the group. For this reason, if the guidance information should be delivered to a certain user, but the attribute information that classifies the user into a predetermined group is not associated with the group, the group is not associated with the group. It is not possible to classify users into a given group. As a result, the guidance information cannot be delivered to the user, and the guidance destination cannot be guided.

The present application has been made in view of the above, and an object of the present application is to classify users more appropriately.

The classification device according to the present application includes a calculation unit that calculates an index value based on the possibility that the user who performed the action will reach a predetermined action for each action of the user, and an action performed by the user to be classified. It is characterized by having a classification unit that classifies the user based on the index value calculated for.

According to one aspect of the embodiment, the users can be classified more appropriately.

FIG. 1 is a diagram showing an example of a classification process executed by the classification device according to the embodiment. FIG. 2 is a diagram showing a configuration example of the classification device according to the embodiment. FIG. 3 is a diagram showing an example of information registered in the query log database according to the embodiment. FIG. 4 is a diagram showing an example of information registered in the arrival query database according to the embodiment. FIG. 5 is a diagram showing an example of information registered in the group database according to the embodiment. FIG. 6 is a diagram showing an example of information registered in the user database according to the embodiment. FIG. 7 is a diagram showing an example of information registered in the content database according to the embodiment. FIG. 8 is a diagram illustrating a concept of a classification process executed by the classification device according to the embodiment. FIG. 9 is a flowchart showing an example of a flow of processing for calculating an index value among the classification processes executed by the classification device according to the embodiment. FIG. 10 is a flowchart showing an example of a flow of processing for classifying users according to an index value of an input query input by a user to be classified among the classification processes executed by the classification device according to the embodiment. is there. FIG. 11 is a diagram showing an example of a hardware configuration.

Hereinafter, a classification device, a classification method, and a mode for carrying out the classification program according to the present application (hereinafter, referred to as “the embodiment”) will be described in detail with reference to the drawings. The classification device, classification method, and classification program according to the present application are not limited by this embodiment. Further, in each of the following embodiments, the same parts are designated by the same reference numerals, and duplicate description is omitted.

[Embodiment]
[1-1. An example of a sorting device]
First, an example of the classification process executed by the classification device will be described with reference to FIG. FIG. 1 is a diagram showing an example of a classification process executed by the classification device according to the embodiment. In FIG. 1, the classification device 10 can communicate with the user terminal 101 used by the user U01 and the user terminal 102 used by the user U02 via a predetermined network N such as the Internet.

User terminals 101 and 102 (hereinafter, may be collectively referred to as "user terminal 100") are smart devices such as smartphones and tablets, and wireless such as 3G (3rd Generation) and LTE (Long Term Evolution). It is a mobile terminal device that can communicate with any server device via a communication network. The user terminal 100 may be an information processing device such as a desktop PC or a notebook PC as well as a smart device.

The classification device 10 is a classification device for classifying users U01 and U02, and is realized by, for example, a server device or a cloud system. For example, the classification device 10 groups each user U01 and U02 based on the possibility of performing a predetermined predetermined action. Then, the classification device 10 selects the content to be distributed to each user U01 and U02 according to the group in which each user U01 and U02 are classified, and distributes the selected content.

For example, the classification device 10 has an act of browsing a predetermined content (for example, a web page, etc.), an act of selecting a predetermined content (for example, clicking or tapping), an act of purchasing a predetermined product or service, bidding, or the like. Etc., preset groups that have different possibilities of reaching a predetermined preset conversion. To give a more specific example, the classification device 10 acquires and browses a web page related to a tour to Hawaii conducted by company # A and a web page related to a tour to Hawaii conducted by company # B. Is the predetermined conversion. Then, the classification device 10 classifies the users U01 and U02 into one of the groups. For example, the classification device 10 classifies each user U01 and U02 into a reverse funnel type group based on the stage until each user U01 and U02 reaches a predetermined conversion, and based on the classification result, the content Deliver.

Here, consider the act of browsing the tour application page as a predetermined conversion. Users who are classified into groups that are likely to lead to such behavior are more likely to actually view the tour application page, but flow into the application page of a competitor who conducts a similar tour for comparison. There is a high probability that it will be done. As a result, in a group that is likely to reach a predetermined conversion, competition between competitors occurs, so that the value of CVR (Conversion Rate) when the advertisement is delivered may decrease. On the other hand, users who are classified into groups that are unlikely to view the tour application page will generally have a lower probability of actually viewing the tour application page, but will conduct similar tours for comparison. Since the act of browsing the application page of the competitor is not performed, the value of CVR when the advertisement is delivered becomes high. Therefore, when guiding each user U01 and U02 to the application page of the tour conducted by a predetermined company, the content and distribution mode of the advertisement are switched according to the group in which each user U01 and U02 are classified. It is considered better.

However, in the prior art, since each user U01 and U02 are grouped based on the relationship between the preset attribute information and the group, it may not be possible to perform appropriate grouping. For example, in the prior art, if the attribute information of the user U01 is not associated with a group that is likely to lead to a predetermined conversion, even if the user U01 is likely to reach a predetermined conversion. , The user U01 is classified into a group that is unlikely to reach a predetermined conversion.

[1-2. An example of classification processing]
Therefore, the classification device 10 executes the following classification process. First, the classification device 10 calculates an index value based on the possibility that the user who has performed the action will reach a predetermined action for each action of the user. Then, the classification device 10 classifies the users based on the index values calculated for the actions performed by the users to be classified.

[1-2-1. About behavior]
Here, the index value based on the possibility of reaching a predetermined action is a value based on the possibility that the user executes a preset action. To give a more specific example, the classification device 10 has a possibility that a user browses a website in a predetermined field, a possibility of browsing a predetermined website, and a purchase of a predetermined product or service. An index value is calculated based on the possibility of selecting, viewing, purchasing, etc. a predetermined content such as a possibility, music content, video content, or the like. That is, the classification device 10 calculates an index value based on the possibility that the user will reach a preset predetermined conversion. In addition, if the classification device 10 calculates an index value based on the possibility that the user who performed the action will reach a predetermined action for each action of the user, the index based on the possibility of reaching an arbitrary action. You may calculate the value.

Further, the user's behavior is, for example, a user's behavior that can be specified on the network, and for each content of the user's behavior, there is a possibility that the user who has performed the behavior will reach a predetermined behavior. Any action can be applied if the index value is calculated based on. For example, the classification device 10 calculates an index value based on the possibility that the user who has performed the web search of the content will reach a predetermined action for each content of the web search by the user. Then, the classification device 10 may classify the user based on the index value calculated for the content of the web search performed by the user to be classified. More specifically, the classification device 10 calculates an index value based on the possibility that the user who entered the search query will reach a predetermined action for each search query entered by the user, and uses the classification target. The user may be classified based on the index value calculated for the search query entered by the user.

Further, for example, the classification device 10 performs a browsing act or a selection act of the content for each content of the browsing act or the selection act by the user, such as the act of browsing the content or the act of selecting a link to the content. Calculate an index value based on the possibility that the user will reach a predetermined action. Then, the classification device 10 may classify the user based on the index value calculated for the content of the browsing act or the selection act performed by the user to be classified. More specifically, the classification device 10 calculates an index value based on the possibility that the user who browses or selects the content will reach a predetermined action for each content browsed or selected by the user, and sets the content as a classification target. The user may be classified based on the index value calculated for the content viewed or selected by the user.

[1-2-2. About index value based on possibility]
Here, the classification device 10 calculates an index value based on the possibility that the user who has performed the action will reach a predetermined action for each action. For example, the classification device 10 calculates an index value for each search query based on the possibility that the user who entered the search query will reach a predetermined action (that is, the possibility of conversion). Here, the classification device 10 calculates an index value based on the possibility that a user who has performed an action for which an index value is calculated (hereinafter, may be referred to as “target action”) will reach a predetermined action. If so, the index value may be calculated by any method.

In addition, the classification device 10 is divided into the content of the action performed by the user who actually reached the predetermined action (hereinafter, referred to as "reaching user") and the content of the action performed by each user. Based on this, an index value based on the possibility that the user will reach a predetermined action is calculated. For example, the classification device 10 may calculate the ratio of the reaching users who performed the target behavior for each target behavior, and calculate an index value based on the calculated ratio. Specifically, the classification device 10 counts the number of all users who have performed the target action and the number of reaching users among the users who have performed the target action. Then, the classification device 10 calculates a value obtained by dividing the number of reaching users among the users who have performed the target action by the number of all users who have performed the target action as an index value.

Further, for example, the classification device 10 calculates a value obtained by dividing the number of times the reaching user has performed the target action for each target action by the number of times each user has performed the target action as an index value. May be good. For example, the classification device 10 acquires a history of search queries (hereinafter, referred to as "query log") input by the user. Further, the classification device 10 acquires a search query (hereinafter, referred to as "arrival query") input by the arrival user who has reached a predetermined conversion. More specifically, the classification device 10 acquires the arrival query input by the arrival user during a predetermined period such as 30 minutes until the user reaches the conversion.

Then, the classification device 10 selects a search query to be calculated for the index value, and counts the number of queries that are the same as the selected search query in the query log and the number of queries that are the same as the selected search query in the reached queries. To do. That is, the classification device 10 counts the number of times each user inputs the search query for which the index value is calculated and the number of times the reaching user inputs the search query.

Further, the classification device 10 divides the number of times the arrival user inputs the search query by the number of times each user inputs the search query for which the index value is calculated, so that the arrival user makes the search query. Calculate the probability of entering. Such a probability can be an index of the possibility that the user who entered the search query will be the reaching user. Therefore, the classification device 10 uses the calculated probability as an index value of the search query. That is, the classification device 10 calculates the probability that the user who has performed a certain target action browses or selects a predetermined content, and the calculated probability is referred to as the reaching user by the user who has performed the target action. The index value is based on the possibility of becoming.

[1-2-3. About classification of users based on index values]
Further, the classification device 10 specifies an index value calculated for the behavior of the user to be classified, and groups the users to be classified based on the specified index value. Specifically, the classification device 10 classifies the users to be classified into one of a plurality of groups having different possibilities of reaching a predetermined behavior. For example, the classification device 10 classifies users into a group that is more likely to reach a predetermined action as the value of the index value calculated for the action performed by the user to be classified is higher. In the following description, the possibility of reaching a predetermined action such as the possibility of reaching a predetermined conversion may be described as an inflow distance, and each group may be described as a group having a different inflow distance.

A more specific example will be described. First, the classification device 10 sets three groups having different probabilities of reaching a predetermined conversion. For example, the classification device 10 includes a group # 1 in which the probability of reaching a predetermined conversion is less than "1"%, and a group # 2 in which the probability of reaching a predetermined conversion is "1"% or more and less than "2"%. Group # 3 with a probability of reaching a predetermined conversion of "2"% or more is set. Further, the classification device 10 identifies the index value having the highest value among the index values calculated for each action performed by the user to be classified. Then, the classification device 10 classifies the users to be classified into the group corresponding to the value of the specified index value. For example, when the value of the specified index value is, for example, "1.3"%, the classification device 10 classifies the users to be classified into the group # 2 corresponding to the index value.

[1-2-4. Providing information based on classification results]
In addition, the classification device 10 provides the user with content according to the classification result. For example, when the classification device 10 receives a request from the user to distribute the content related to the advertisement from the user, the classification device 10 classifies the user U01 based on the index value of the action performed by the user U01. Then, the classification device 10 selects the content corresponding to the group in which the user U01 is classified, and distributes the selected content to the user U01.

More specifically, the classification device 10 holds a plurality of contents in association with each other for each group to be classified. Then, when the classification device 10 receives the request from the user U01, the classification device 10 identifies a plurality of contents corresponding to the group in which the user U01 is classified, and the user U01 is referred to from the specified plurality of contents. The content to be distributed is selected based on the behavior of the user U01. Then, the classification device 10 distributes the selected content to the user U01.

In this way, the classification device 10 calculates an index value based on the possibility that the user who performed the action will reach a predetermined action for each action of the user, and based on the calculated index value, sets the user. Classify. As a result, the classification device 10 classifies each user according to the possibility of performing a predetermined action, so that the users can be classified more appropriately. For example, since the classification device 10 groups the user U01 according to the possibility of reaching a predetermined conversion, the grouping is appropriately performed when the guidance information regarding the predetermined conversion is distributed to the user U01. be able to.

[1-3. About an example of classification processing]
Next, an example of the classification process executed by the classification device 10 will be described with reference to FIG. In the following description, a predetermined target action such as conversion is described as a “target”. In addition, in the following explanation, an example of processing targeting browsing, selection, etc. of the introduction page and application page of each company related to the tour to Hawaii will be described. Further, in the following description, as an example of the classification process, the classification device 10 calculates an index value based on the possibility of reaching the target, that is, the inflow distance to the target, and classifies each search query entered by the user. An example of a process of classifying users based on a search query entered by a target user and a corresponding inflow distance will be described.

First, the classification device 01 estimates the correspondence between the inflow distance to the target and the search query based on the appearance probability of the query input by the arriving user (step S1). That is, the classification device 10 calculates an index value based on the possibility of reaching the target for each search query. For example, the classification device 10 acquires a query log, which is a history of search queries input by various users, and registers it in the query log database 31. Further, the classification device 10 acquires the arrival query, which is the history of the search query input by the arrival user, and registers it in the arrival query database 32 within 30 minutes until the arrival user reaches the target. Then, the classification device 10 calculates the appearance probability of the arrival query (step S2). More specifically, the classification device 10 calculates the appearance probability as the value obtained by dividing the number of times the arrival query is input by the arrival user for each arrival query by the number of times each user inputs the arrival query. .. That is, the classification device 10 calculates a value based on the probability that the user who input the arrival query becomes the arrival user for each arrival query.

Then, the classification device 10 estimates the correspondence between the inflow distance and the query based on the appearance probability (step S3). More specifically, the classification device 10 divides each arrival query into groups having different inflow distances based on the value of the probability calculated for each arrival query. For example, the classification device 10 associates arrival queries such as “Guam”, “Company # A”, “Company # B”, etc., which have a relatively low calculated probability, with “Group # 1” having the longest inflow distance to the target. Further, for example, the classification device 10 targets arrival queries such as "Hawaii", "Waikiki Hotel", "Pack Tour Company #A", etc., for which the calculated probability is medium (for example, 1% or more and less than 3%). Corresponds to "Group # 2", which has the second longest inflow distance. Further, for example, the classification device 10 associates an arrival query such as "company #A Hawaii" with a high calculated probability (for example, 3% or more) with "group # 3" having the closest inflow distance to the target.

Subsequently, the classification device 10 receives a request from the users U01 and U02 (step S4). The timing at which the classification device 10 receives requests from the users U01 and U02 does not have to be simultaneous. For example, the classification device 10 receives a distribution request for content related to an advertisement to be displayed together with a web page from the user terminal 100. In such a case, the classification device 10 groups the users into groups having different inflow distances to the target according to the input queries of the users U01 and U02 (step S5).

For example, the classification device 10 acquires an input query which is a search query input by the user U01 within a predetermined time (for example, during 30 minutes until the time of classification). For example, in the example shown in FIG. 1, the classification device 10 acquires the input query "company #A Hawaii", "gourmet site", "blog", "Waikiki hotel", "soccer", etc. of the user U01. In such a case, the classification device 10 distributes the user U01 to the group whose inflow distance is closest to the target among the groups associated with the input query of the user U01. For example, the classification device 10 distributes the user U01 to the group # 3 to which the input query “Company # A Hawaii” is associated.

On the other hand, in the example shown in FIG. 1, the classification device 10 acquires the input queries "soccer", "gourmet", "baseball", "Guam", "blog", etc. of the user U02. In such a case, the classification device 10 distributes the user U02 to the group whose inflow distance is closest to the target among the groups associated with the input query of the user U02. For example, the classification device 10 distributes the user U02 to the group # 1 to which the input query "Guam" is associated.

After that, the classification device 10 delivers the advertisement corresponding to the group # 3 to the user U01 (step S6). For example, the classification device 10 delivers to the user U01 the content selected from the viewpoint of further enhancing the CVR among the advertisements related to the target. On the other hand, the classification device 10 delivers the advertisement corresponding to the group # 1 to the user U02 (step S7). For example, the classification device 10 distributes the content selected from the viewpoint of notifying the user U02 of the content of the target by targeting the user U02 with various advertisements related to the target.

For example, the classification device 10 may distribute the content in different distribution modes between the users of the group # 1 and the group # 2 and the users of the group # 3. For example, the classification device 10 distributes the content to the users of the group # 1 and the group # 2 as an advertisement or the like placed in the web content. On the other hand, the classification device 10 may push-deliver contents such as advertisements to users of group # 3 in the form of e-mails, messages, and the like.

[2. Classification device configuration]
Hereinafter, an example of the functional configuration of the classification device 10 that realizes the above-mentioned classification process will be described. FIG. 2 is a diagram showing a configuration example of the classification device according to the embodiment. As shown in FIG. 2, the classification device 10 includes a communication unit 20, a storage unit 30, and a control unit 40.

The communication unit 20 is realized by, for example, a NIC (Network Interface Card) or the like. Then, the communication unit 20 is connected to the network N by wire or wirelessly, and transmits / receives information to / from the user terminal 100, the search log server 200, and the action log server 300. The search log server 200 is a server device that manages a log of search queries entered by each user, that is, a query log. The action log server 300 is a server device that manages information indicating the content browsed or selected by the user.

The storage unit 30 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. Further, the storage unit 30 stores the query log database 31, the arrival query database 32, the group database 33, the user database 34, and the content database 35 (hereinafter, may be collectively referred to as “each database 31 to 35”). Hereinafter, an example of information registered in each database 31 to 35 will be described with reference to FIGS. 3 to 7.

In the query log database 31, a query log, which is a history of search queries entered by various users, is registered. For example, FIG. 3 is a diagram showing an example of information registered in the query log database according to the embodiment. As shown in FIG. 3, a query log, which is a history of search queries input by various users within a predetermined period, is registered in the query log database 31. For example, in the example shown in FIG. 3, various search queries such as "company # A", "Hawaii", and "SNS # A" are registered in the query log database 31.

In the arrival query database 32, a user who has reached a predetermined action and a arrival query which is a search query input by the reaching user during a predetermined period (for example, 30 minutes) until the predetermined action is registered are registered. To. For example, FIG. 4 is a diagram showing an example of information registered in the arrival query database according to the embodiment. As shown in FIG. 4, information having items such as "arrival query", "probability", and "flag" is registered in the arrival query database 32.

Here, the "arrival query" is an arrival query input by the arrival user. Further, the "probability" is the appearance probability of the associated "arrival query". Further, the "flag" is information indicating whether or not the content of the associated "arrival query" is related to a preset sub-item. For example, in the example shown in FIG. 4, various sub-items such as "Hawaii", "proximity", "other overseas", "competitor", "your company", "price", "travel", etc. It is included. In addition to the items shown in FIG. 4, various items may be registered in the arrival query database 32.

For example, in the example shown in FIG. 4, the arrival query "Company #A Hawaii" and the probability "3.0" are supported by flags indicating that they correspond to the sub-items "Hawaii", "Your company", and "Travel". It is registered with. For such information, the probability of appearance calculated for the arrival query "Company # A Hawaii" is "3.0"%, and the arrival query "Company # A Hawaii" is "Hawaii", "Your company", "Travel". Indicates that it is an arrival query related to. In the example shown in FIG. 4, an example of an arrival query corresponding to a certain target has been described, but in the arrival query database 32, information having the same configuration as the information shown in FIG. 4 is registered for each target. It is assumed that it has been done.

In the group database 33, information on a group that classifies users is registered for each target. For example, FIG. 5 is a diagram showing an example of information registered in the group database according to the embodiment. In the example shown in FIG. 5, information such as "target ID (Identifier)", "group ID", "condition (probability)", and "arrival query" is registered in the group database 33.

Here, the "target ID" is information for identifying a target such as a predetermined action or conversion. The "group ID" is information for identifying groups having different inflow distances to the corresponding targets. Further, the "condition (probability)" is information indicating the conditions used when classifying the content reached to the associated group and the users to be classified, and more specifically, the information is classified into the group. It is information indicating the range of the appearance probability of the search query input by the user.

For example, in the example shown in FIG. 5, the target ID “target # 1”, the group ID “group # 1”, the condition (probability) “1% or less”, and the arrival query “Guam” are registered in association with each other. Such information corresponds to, for example, the arrival query in which the probability of reaching the target indicated by the target ID "target # 1" is the condition (probability) "1% or less" among the arrival queries corresponds to the group ID "group # 1". Indicates that it is attached. Further, such information indicates that the arrival query "Guam" is associated with the group indicated by the group ID "Group # 1" among the groups corresponding to the target ID "Target # 1". ..

The group to which each user belongs is registered in the user database 34. For example, FIG. 6 is a diagram showing an example of information registered in the user database according to the embodiment. In the example shown in FIG. 6, items such as "user ID", "target ID", and "group ID" are registered in the user database 34 in association with each other. In addition to the information shown in FIG. 6, arbitrary information about the user may be registered in the user database 34.

Here, the "user ID" is information for identifying a user. For example, in the example shown in FIG. 6, the user ID "user # 1", the target ID "target # 1", and the group ID "group # 2" are registered in association with each other in the user database 34. Such information indicates that, for example, among the groups set for the target ID "target # 1", the user indicated by the user ID "user # 1" is classified into the group indicated by the group ID "group # 2". Is shown.

In the content database 35, contents to be distributed according to the classification result of the user are registered. More specifically, in the content database 35, content related to advertisements is registered as content to be distributed to users. For example, FIG. 7 is a diagram showing an example of information registered in the content database according to the embodiment. In the example shown in FIG. 7, items such as "advertisement ID", "advertisement data", "target ID", and "group ID" are registered in the content database 35 in association with each other.

Here, the "advertisement ID" is information for identifying the advertisement delivered to the user. Further, the "advertisement data" is the data of the advertisement delivered to the user. For example, in the example shown in FIG. 7, the content database 35 corresponds to the advertisement ID “advertisement # 1”, the advertisement data “advertisement data # 1”, the target ID “target # 1”, and the group ID “group # 1”. It is registered with. Such information is provided to the users classified into the group indicated by the group ID "Group # 1" among the groups set for the target ID "Target # 1" in the advertisement indicated by the advertisement ID "Advertisement # 1". Indicates that the advertisement data "advertisement data # 1" will be delivered.

Returning to FIG. 2, the explanation will be continued. The control unit 40 is a controller, and for example, various programs stored in a storage device inside the classification device 10 by a processor such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit) store a RAM or the like. It is realized by being executed as a work area. Further, the control unit 40 is a controller, and may be realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 2, the control unit 40 includes an acquisition unit 41, a calculation unit 42, a specific unit 43, a setting unit 44, a reception unit 45, a classification unit 46, and a provision unit 47.

The acquisition unit 41 acquires the query log and the arrival query. For example, the acquisition unit 41 acquires a search query from the search log server 200 at a predetermined time interval before the predetermined period elapses after the input. Then, the acquisition unit 41 registers the acquired search query as a query log in the query log database 31.

In addition, the acquisition unit 41 acquires information indicating a user who has performed a predetermined action, that is, a reaching user, from the action log server 300 at a predetermined time interval. Then, the acquisition unit 41 acquires the arrival query, which is the search query input by the arrival user, from the search log server 200, and registers the acquired arrival query in the arrival query database 32. In addition, the acquisition unit 41 acquires a search query input within a predetermined period, such as a search query input during 30 minutes until the arrival user performs a predetermined action, as an arrival query among the search queries input by the reaching user. You may.

Further, the acquisition unit 41 stores the flag of the sub-item indicating the attribute of the arrival query in the arrival query database 32 by using a predetermined character analysis technique or the like. For example, when the acquired arrival query is "Guam", the acquisition unit 41 sets a flag of a small item such as "other overseas" or "travel system" in which the arrival query is associated with "Guam". Update to "1".

The calculation unit 42 calculates an index value based on the possibility that the user who performed the action will reach a predetermined action for each action of the user. Specifically, the calculation unit 42 calculates the value of the appearance probability for each arrival query, and registers the calculated value of the appearance probability in the arrival query database 32. For example, the calculation unit 42 calculates, as an index value, a value obtained by dividing the number of times the reaching user has performed the action for each user's action by the number of times each user has performed the action. More specifically, the calculation unit 42 calculates an index value based on the possibility that the user who performed the action will reach the predetermined action for each action performed by the reaching user during the predetermined period. In addition, the calculation unit 42 calculates the ratio of users who have achieved a predetermined behavior among the users who have performed the behavior for each behavior of the user, and calculates an index value based on the calculated ratio.

Hereinafter, an example of the processing executed by the calculation unit 42 will be described. For example, the calculation unit 42 calculates an index value based on the possibility that the user who entered the search query will reach a predetermined action for each search query entered by the user as the action of the user. More specifically, the calculation unit 42 refers to the arrival query database 32 and selects one arrival query for which the appearance probability has not been calculated as a processing target. Subsequently, the calculation unit 42 refers to the query log database 31 and counts the number of times (hereinafter, referred to as “parameter”) that the same search query as the arrival query selected as the processing target is included in the query log. .. Further, the calculation unit 42 identifies the arrival query input by the arrival user from the query log, and the number of arrival queries selected as the processing target in the specified specific query (hereinafter referred to as "number of children"). Describe.) Is counted. The calculation unit 42 may acquire the number of times each arrival user inputs the arrival query to be processed from the search log server 200 or the like as an index.

Then, the calculation unit 42 calculates the value obtained by dividing the counted number of children by the population parameter as the appearance probability of the arrival query selected as the predetermined target. That is, the calculation unit 42 calculates an index value based on the possibility that the user who has input the selected arrival query will reach a predetermined action or conversion. After that, the calculation unit 42 registers the calculated appearance probability in the arrival query database 32. The calculation unit 42 calculates the appearance probability for all the arrival queries by repeatedly executing the above-mentioned processing.

Here, the calculation unit 42 may execute another process for each action of the user as long as it calculates an index value based on the possibility that the user who performed the action will reach a predetermined action. .. For example, the calculation unit 42 may calculate an index value based on the possibility that the user who browses or selects the content will reach a predetermined behavior for each content browsed or selected by the user as the behavior of the user. .. More specifically, the calculation unit 42 divides the number of times the reaching user has viewed or selected the content by the number of times each user has viewed or selected the content, and calculates the value of the content. It may be calculated as an index value. Further, the calculation unit 42 may calculate the index value from the calculated value in consideration of various weightings.

The identification unit 43 identifies the behavior of the user corresponding to a plurality of groups having different inflow distances based on the index value. More specifically, the identification unit 43 specifies a search query input by a user classified into each group corresponding to a predetermined target.

For example, the specific unit 43 refers to the arrival query database 32 and extracts the arrival query corresponding to the target to be processed. Subsequently, the identification unit 43 specifies the appearance probability of the extracted arrival query. Then, the specific unit 43 specifies the group corresponding to the extracted arrival query according to the value of the specified appearance probability.

For example, when the arrival query shown in FIG. 4 is registered in the arrival query database 32 as the arrival query corresponding to the target ID “target # 1”, the specific unit 43 has an appearance probability of the arrival query “Company # A Hawaii”. Identify "3.0"%. In such a case, the specific unit 43 specifies a group in which the value of the item “condition (probability)” satisfies the appearance probability “3.0”% among the groups corresponding to the target ID “target # 1”. For example, when the condition shown in FIG. 5 is registered in the group database 33, the occurrence probability of the arrival query “Company # A Hawaii” matches the group condition indicated by the group ID “Group # 3” in the specific unit 43. Then it is determined.

The setting unit 44 sets the behavior of the user classified into the group for a plurality of groups having different inflow distances. For example, when the setting unit 44 determines by the specific unit 43 that the appearance probability of the arrival query "company # A Hawaii" matches the conditions of the group indicated by the group ID "group # 3", the group ID "group" The arrival query "Company # A Hawaii" is linked to the group indicated by "# 3". For example, the setting unit 44 associates the target ID “target # 1” with the group ID “group # 3” with the arrival query “company # A Hawaii” and registers them in the group database 33.

The reception unit 45 receives various requests from the user terminal 100. In such a case, the reception unit 45 identifies the user who uses the user terminal 100 as the request source, and notifies the classification unit 46 of the specified user.

The classification unit 46 classifies users based on the index value calculated for the user who is the request source, that is, the action performed by the user to be classified. Specifically, the classification unit 46 classifies the users to be classified into one of a plurality of groups having different possibilities of reaching a predetermined behavior. More specifically, the classification unit 46 classifies users into a group that is more likely to reach a predetermined action as the value of the index value calculated for the action performed by the user to be classified is higher.

For example, the classification unit 46 classifies users based on the index value calculated for the search query input by the user to be classified. Further, in another example, the classification unit 46 classifies the users based on the index value calculated for the content viewed or selected by the user to be classified.

Hereinafter, an example of the processing executed by the classification unit 46 will be described. First, the classification unit 46 refers to the user database 34 and determines whether or not the group corresponding to each target is associated with the user ID indicating the user notified from the reception unit 45. That is, the classification unit 46 determines whether or not the user has already been classified. Then, the classification unit 46 requests the providing unit 47 to provide information when the users are classified.

On the other hand, the classification unit 46 executes the following processing when the users are not classified. First, the classification unit 46 acquires an input query input by the user to be classified within a predetermined period. Then, the classification unit 46 refers to the group database 33 and identifies the group having the closest inflow distance to the target among the groups associated with the acquired input query for each target. For example, when the user has input both the input query "Guam" and the input query "Company # A Hawaii", the classification unit 46 inputs the target ID "Target # 1" from the corresponding groups. The group ID "Group # 1" associated with the query "Guam" and the group ID "Group # 3" associated with the input query "Company # A Hawaii" are specified. Then, the classification unit 46 indicates a group among the specified group ID “group # 1” and the group ID “group # 3” that has a closer inflow distance to the target indicated by the target ID “target # 1”. The ID, that is, the group ID "group # 3" is specified. For example, the classification unit 46 sets the group ID of the group having the higher value of the associated "condition (probability)" among the specified group IDs as the group ID indicating the group having the closer inflow distance to the target. ..

In such a case, the classification unit 46 registers the user ID of the user, the target ID “target # 1”, and the group ID “group # 3” in the user database 34 in association with each other. Further, the classification unit 46 executes the same processing for other target IDs to identify the group to which the users to be classified belong and classify the users. Then, the classification unit 46 also requests the provision unit 47 to provide information when the users are classified for each target ID.

The providing unit 47 provides the user with the content according to the classification result. More specifically, the providing unit 47 provides the content selected according to the action taken by the user among the contents according to the classification result. For example, the providing unit 47 identifies the target ID and the group ID associated with the user who is the request source from the user database 34. Next, the providing unit 47 refers to the content database 35 and identifies the advertisement associated with the pair of the target ID and the group ID specified from the user database 34. Then, the providing unit 47 delivers the specified advertisement to the user who requested it.

In addition, the providing unit 47 can select any targeting technique as to which of the specified advertisements is to be delivered. For example, when the providing unit 47 identifies a plurality of advertisements associated with the same target ID and group ID pair, the provided unit 47 of the specified advertisement is based on the input query input by the requesting user. You may further select the advertisement to be delivered from the list and deliver the selected advertisement to the requesting user. For example, the providing unit 47 identifies a sub-item related to the input query input by the requesting user based on the flag registered in the arrival query database 32, and delivers according to the specified sub-item. You may select the target advertisement. That is, the providing unit 47 may change the content to be delivered according to the appearance probability of the input query input by the user of the request source and the contents of the sub-items corresponding to the input query.

[3. Concept of classification process executed by the classification device]
Next, the concept of the classification process executed by the classification device 10 will be described with reference to FIG. FIG. 8 is a diagram illustrating a concept of a classification process executed by the classification device according to the embodiment. In the example shown in FIG. 8, as an example of the classification process executed by the classification device 10, the index value (that is, the inflow distance) to the target is calculated for each search query, and the search input by the user to be classified is performed. An example of classification processing that classifies users according to the index value of the query is described.

First, as shown in FIG. 8A, the classification device 10 calculates the probability that the user who has input each search query will reach a predetermined action. For example, in the example shown in FIG. 8, the arrival queries entered by each arrival user U10 include "Guam", "Blog", "Company # A", "Waikiki Hotel", "Pack Tour Company # A", and the like. The query exists. In such a case, the classification device 10 determines the number of times that the arrival user U10 inputs each arrival query within a predetermined period until the arrival at the target (that is, within a predetermined period until the predetermined action is performed). , Calculated for each arrival query. On the other hand, the classification device 10 calculates the number of times that all the users U20 input each arrival query within a predetermined period until the arrival at the target for each arrival query. Then, the classification device 10 uses a value obtained by dividing the number of times the arrival user U10 inputs the arrival query by the number of times all the users U20 input the arrival query as the appearance probability for each arrival query.

Here, as shown in FIG. 8B, a general search query that is not related to conversion, such as "blog", that is, an arrival query that has little relevance to the target is a search query that can be generally input. Therefore, it is considered that all users U20 are also frequently input. Therefore, the value of the appearance probability of a general search query that is not related to conversion such as "blog" is considered to be a low value to some extent. On the other hand, as shown in FIG. 8C, it is considered that the arrival query having a significant related example to the conversion such as "Waikiki Hotel" is input more frequently by the arriving user U01 than by all the users U20. Be done. Therefore, the value of the appearance probability of the arrival query having a significant related example to the conversion such as "Waikiki Hotel" is considered to be a high value to some extent.

Therefore, the classification device 10 identifies the relationship between the arrival query and the target by calculating the appearance probability for each arrival query. Then, as shown in FIG. 8D, the classification device 10 classifies the users based on the specific result. For example, the classification device 10 is input by the user U01 to be classified during a predetermined period until the processing (reference time) for classifying the user U01, as shown in FIG. 8 (E). Get an input query. For example, the classification device 10 acquires "baseball," "Waikiki hotel," "soccer," "blog," and the like.

Here, it is considered that the input query "baseball" shown in FIG. 8 (F) is not so related to the target. On the other hand, the user U01 inputs an input query such as "Waikiki Hotel" as shown in FIG. 8 (G). Therefore, the classification device 10 determines the user U01 based on the value of the appearance probability calculated for the "Waikiki hotel", that is, the index value of the possibility that the user who inputs "Waikiki hotel" becomes the reaching user U10. Classify. For example, the classification device 10 classifies the user U01 into group # 2 in which the inflow distance to the target is close to some extent.

In this way, the classification device 10 classifies the user U01 based on the index value of the possibility that the user who input the same search query as the input query input by the user U01 becomes the reaching user U10. Therefore, the classification device 10 can more appropriately classify the user U01.

[4. An example of the flow of classification processing]
Subsequently, an example of the flow of the classification process executed by the classification device 10 will be described with reference to FIGS. 9 and 10. FIG. 9 is a flowchart showing an example of a flow of processing for calculating an index value among the classification processes executed by the classification device according to the embodiment. Further, FIG. 10 shows an example of a flow of a process of classifying users according to an index value of an input query input by a user to be classified among the classification processes executed by the classification device according to the embodiment. It is a flowchart.

First, an example of the flow of processing for calculating the index value will be described with reference to FIG. For example, the classification device 10 acquires a query log (step S101). Subsequently, the classification device 10 acquires an arrival log, which is a log of arrival queries for a predetermined target (step S102). Then, the classification device 10 uses the query log and the arrival log to calculate an index value based on the possibility that the user who input the arrival query will reach a predetermined target for each arrival query (step S103). Then, the classification device 10 associates each arrival query with each group having a different inflow distance to the target based on the calculated index value (step S104), and ends the process.

Next, an example of the flow of processing for classifying users will be described with reference to FIG. For example, the classification device 10 determines whether or not the request has been received from the user (step S201), and if it has received the request (step S201: Yes), the request source user has been classified into any group. Whether or not it is determined (step S202). Then, when the request source user is not classified into a group (step S202: No), the classification device 10 identifies an input query input by the user within a predetermined time (step S203), and identifies the specified input. The users are classified into each group according to the index value of the query (step S204).

Subsequently, the classification device 10 selects an advertisement according to the group that classified the users (step S205), delivers the selected advertisement (step S206), and ends the process. On the other hand, if the requesting user has been classified into any group (step S202: Yes), the classification device 10 executes step S205. On the other hand, if the classification device 10 has not received the request from the user (step S201: No), the classification device 10 waits until it is received.

[5. Modification example]
In the above, an example of the classification process by the classification device 10 has been described. However, the embodiment is not limited to this. Hereinafter, variations of the classification process executed by the classification device 10 will be described.

[5-1. About classification]
The classification device 10 described above classifies users into a plurality of groups having different inflow distances to preset targets based on the index value, and delivers advertisements according to the classification results. Here, the classification device 10 may classify each user into an arbitrary number of groups. For example, the classification device 10 may classify the users into two groups having different inflow distances to the target, and classifies the users into an arbitrary number of groups set according to the wishes of the distribution source of the guidance information. May be good. Further, in the above-described example, the classification device 10 calculates an index value for each search query, sets the correspondence between the search query and each group according to the calculated index value, and sets the users U01 and U02 as users. It was classified into a group associated with the input query input by U01 and U02. However, the embodiment is not limited to this. For example, the classification device 10 may calculate an index value for each search query and classify the users U01 and U02 based on the index value of the same search query as the input query input by the users U01 and U02. Good. That is, the classification device 10 does not have to associate the group with the search query as long as the users can be classified.

[5-2. About processing after classification]
Further, the classification device 10 may analyze users in units of groups that classify users. For example, consider browsing a travel application page run by a different company as a conversion, and classifying users into groups with different inflow distances to the conversion. In such groups, not only the possibility of conversion, but also the rate of competition and the rate of loss that does not result in conversion differ from group to group.

Therefore, the classification device 10 may count the rate of competition, the rate of loss, etc. for each group, and provide the counted content to the advertiser or the like as a report. Further, the classification device 10 may select the content to be distributed according to the content counted for each group, or may change the content content and the distribution mode. For example, the classification device 10 delivers various advertisements for the purpose of enclosing a group with a long inflow distance but little competition, and a discount price or a sale for a group with a short inflow distance but a lot of competition. Processing such as presenting the price may be performed. Further, the classification device 10 delivers an advertisement displayed together with the web page to the users of the group having a long inflow distance, and in the form of direct mail or the like to the users of the group having a short inflow distance. , You may push the advertisement.

Further, the classification device 10 may measure the conversion depth for each group and execute various processes according to the measured conversion depth. For example, when the classification device 10 targets the purchase of a certain product, the classification device 10 calculates the ratio of whether each user purchased the product or browsed the web page of the product for each group. That is, the classification device 10 calculates the rate at which actions with different conversion depths are performed for each group. Then, the classification device 10 may sort out advertisements based on the ratio calculated for each group.

[5-3. About target setting]
The classification device 10 can set an arbitrary action as a target. Here, from the viewpoint of increasing the number of user behaviors (for example, search queries) for which the index value is calculated, that is, increasing the number of samples, the behaviors set as targets have the same purpose. It is better to include as much of the same kind of behavior as possible to meet. For example, when targeting browsing of the tour application page for Hawaii, in order to increase the number of search queries for which index values are calculated, not only tours of a given company but also tours provided by competitors Browsing the application page should also be included in the target. As a result, the classification device 10 can increase the number of actions for which the index value is calculated, so that the classification accuracy of the user can be further improved. The target when the classification device 10 calculates the index value for each search query (that is, for each user behavior) is various information to be distributed to the users U01 and U02 after classifying the users U01 and U02. It may be a related target or an unrelated target.

Further, the classification device 10 may change the period in which the search query and the arrival query used for calculating the index value are input according to the target. For example, when purchasing a condominium or a high-priced product, it is considered that it takes a long time for the user to convert. Therefore, when the classification device 10 targets real estate-related acts, purchase acts of high-priced products, etc., the period in which the search query and the arrival query used for calculating the index value are input is set to a longer period. You may. For example, the classification device 10 may use the search query input by the reaching user as the reaching query during the one month until the reaching user reaches the conversion.

Further, in the classification device 10, among the search queries input by each user, the period during which the search query used for calculating the index value was input and the arrival query among the search queries input by the reaching user were input. The period may be set to a different value. For example, the classification device 10 may use the search query input in the past one month as a query log and the search query input in one hour until the arrival user reaches the conversion as the arrival query.
[5-4. Device configuration〕
The classification device 10 may be communicably connected to any number of user terminals 100. Further, the classification device 10 may be realized by a front-end server that exchanges information with the user terminal 100 and a back-end server that selects input items. In such a case, the acquisition unit 41, the calculation unit 42, the specific unit 43, and the setting unit 44 shown in FIG. 2 are arranged on the back-end server, and the reception unit 45, the classification unit 46, and the provision unit are provided on the front-end server. The part 47 and the like will be arranged. Further, each database 31 to 35 stored in the storage unit 30 may be stored not inside the classification device 10 but in, for example, an external storage server or the like.

[5-5. About each database item]
In the above description, the information and information items registered in each database 31 to 35 are merely examples, and any information and items can be registered as long as the above-mentioned classification process can be appropriately performed. May be done.

[5-6. Others]
Further, among the processes described in the above-described embodiment, all or a part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed can be performed. All or part of it can be done automatically by a known method. In addition, the processing procedure, specific name, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically distributed / physically in arbitrary units according to various loads and usage conditions. Can be integrated and configured.

In addition, the above-described embodiments can be appropriately combined as long as the processing contents do not contradict each other.

[6. program〕
Further, the classification device 10 according to the above-described embodiment is realized by, for example, a computer 1000 having a configuration as shown in FIG. FIG. 11 is a diagram showing an example of a hardware configuration. The computer 1000 is connected to the output device 1010 and the input device 1020, and the arithmetic unit 1030, the primary storage device 1040, the secondary storage device 1050, the output IF (Interface) 1060, the input IF 1070, and the network IF 1080 are connected by the bus 1090. Has.

The arithmetic unit 1030 operates based on a program stored in the primary storage device 1040 or the secondary storage device 1050, a program read from the input device 1020, or the like, and executes various processes. The primary storage device 1040 is a memory device that temporarily stores data used by the arithmetic unit 1030 for various calculations, such as a RAM. Further, the secondary storage device 1050 is a storage device in which data used by the arithmetic unit 1030 for various calculations and various databases are registered, and is realized by a ROM (Read Only Memory), an HDD, a flash memory, or the like.

The output IF 1060 is an interface for transmitting information to be output to an output device 1010 that outputs various information such as a monitor and a printer. For example, USB (Universal Serial Bus), DVI (Digital Visual Interface), and the like. It is realized by a connector of a standard such as HDMI (registered trademark) (High Definition Multimedia Interface). Further, the input IF 1070 is an interface for receiving information from various input devices 1020 such as a mouse, a keyboard, and a scanner, and is realized by, for example, USB.

The input device 1020 includes, for example, an optical recording medium such as a CD (Compact Disc), a DVD (Digital Versatile Disc), or a PD (Phase change rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), or a tape. It may be a device that reads information from a medium, a magnetic recording medium, a semiconductor memory, or the like. Further, the input device 1020 may be an external storage medium such as a USB memory.

The network IF1080 receives data from another device via the network N and sends it to the arithmetic unit 1030, and also transmits the data generated by the arithmetic unit 1030 to the other device via the network N.

The arithmetic unit 1030 controls the output device 1010 and the input device 1020 via the output IF 1060 and the input IF 1070. For example, the arithmetic unit 1030 loads a program from the input device 1020 or the secondary storage device 1050 onto the primary storage device 1040, and executes the loaded program.

For example, when the computer 1000 functions as the classification device 10, the arithmetic unit 1030 of the computer 1000 realizes the function of the control unit 40 by executing the program loaded on the primary storage device 1040.

[7. effect〕
As described above, the classification device 10 calculates an index value based on the possibility that the user who has performed the action will reach a predetermined action for each action of the user. Then, the classification device 10 classifies the user based on the index value calculated for the action performed by the user to be classified.

That is, the classification device 10 calculates an index value based on the possibility that the user who actually performed the action reaches the target for each user's action, and the classification device 10 calculates the index value based on the possibility that the user who actually performed the action reaches the target. Users are classified based on the calculated index value. In this way, the classification device 10 calculates the index value based on the actually performed action, not based on the preset correspondence relationship, and classifies according to the calculated index value. As a result, the classification device 10 can classify the users based on the relationship between the actually performed action and the target, so that the classification accuracy of the users can be improved.

In addition, the classification device 10 calculates an index value based on the possibility that the user who input the search query will reach a predetermined behavior for each search query input by the user as the behavior of the user. Then, the classification device 10 classifies the user based on the index value calculated for the search query input by the user to be classified. Therefore, the classification device 10 can appropriately classify the users based on the search query input by the users.

In addition, the classification device 10 calculates an index value based on the possibility that the user who browses or selects the content will reach a predetermined behavior for each content browsed or selected by the user as the behavior of the user. Then, the classification device 10 classifies the user based on the index value calculated for the content viewed or selected by the user to be classified. Therefore, the classification device 10 can appropriately classify the users by using the contents viewed or selected by the users.

In addition, the classification device 10 calculates the ratio of users who have achieved a predetermined behavior among the users who have performed the behavior for each behavior of the user, and calculates an index value based on the calculated ratio. Further, the classification device 10 calculates, as an index value, a value obtained by dividing the number of times the reaching user has performed the action for each user's action by the number of times each user has performed the action. Further, the classification device 10 identifies an action performed by the reaching user during a predetermined period, and for each specified action, sets an index value based on the possibility that the user who performed the action will reach the predetermined action. calculate. In this way, the classification device 10 calculates the index value based on the ratio of the reaching user performing the action for which the index value is calculated. As a result, the classification device 10 can calculate the value of the index value based on the actual measurement, so that the classification accuracy of the user can be further improved.

Further, the classification device 10 calculates, as an index value, an index value based on the possibility that the user browses the predetermined content, or an index value based on the possibility that the user selects the predetermined content. Therefore, the classification device 10 can group users based on the inflow distance to the conversion for viewing or selecting a predetermined content.

In addition, the classification device 10 classifies the users to be classified into any of a plurality of groups having different possibilities of reaching a predetermined behavior. More specifically, the classification device 10 classifies the user into a group that is more likely to reach a predetermined action as the value of the index value calculated for the action performed by the user to be classified is higher. .. As a result, the classification device 10 can group users according to the inflow distance to the target.

In addition, the classification device 10 provides the user with content according to the classification result. For example, the classification device 10 provides the content selected according to the action taken by the user among the contents according to the classification result. In this way, the classification device 10 distributes the content according to the group in which the user is classified, not the action itself such as the search query input by the user. Therefore, for example, the classification device 10 can realize the distribution of contents to the user according to the relationship between the user's behavior and the conversion and the inflow distance to the target.

Although some of the embodiments of the present application have been described in detail with reference to the drawings, these are examples, and various modifications are made based on the knowledge of those skilled in the art, including the embodiments described in the disclosure column of the invention. It is possible to practice the present invention in other improved forms.

Further, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the classification unit can be read as a classification means or a classification circuit.

10 Classification device 20 Communication unit 30 Storage unit 31 Query log database 32 Reach query database 33 Group database 34 User database 35 Content database 40 Control unit 41 Acquisition unit 42 Calculation unit 43 Specific unit 44 Setting unit 45 Reception unit 46 Classification unit 47 Providing unit 100 user terminal

Claims (10)

For each action of the reaching user who has performed a predetermined target action, an index value is calculated based on the value obtained by dividing the number of times the reaching user has performed the action by the number of times each user has performed the action. Calculation part and
Based on the index value calculated by the calculation unit, each of the plurality of groups is divided into one of a plurality of groups in which the user who has performed the action classified into each group has a different possibility of performing the target action. classifying the behavior, based on the classification result of the user to be classified is conducted action classification device characterized in that it has a classifying section for classifying the user. As the behavior of the user, the calculation unit calculates an index value based on the ratio of the user who entered the search query to the predetermined target behavior for each search query entered by the user.
The classification device according to claim 1, wherein the classification unit classifies the user based on an index value calculated for a search query input by the user to be classified. The calculation unit calculates an index value based on the ratio of the user who browses or selects the content to the predetermined target behavior for each content browsed or selected by the user as the behavior of the user. ,
The classification device according to claim 1, wherein the classification unit classifies the user based on an index value calculated for the content viewed or selected by the user to be classified. The calculation unit identifies the actions taken by the user who has reached the predetermined target action during the predetermined period, and for each of the specified actions, the user who has performed the specified action reaches the predetermined target action. The classification device according to any one of claims 1 to 3 , wherein an index value is calculated based on the ratio. The calculation unit is characterized in that, as the index value, an index value based on the possibility that the user browses the predetermined content or an index value based on the ratio of the user selecting the predetermined content is calculated. The classification device according to any one of claims 1 to 4. The classification unit is characterized in that the higher the value of the index value calculated for the behavior performed by the user to be classified, the higher the possibility of reaching the predetermined target behavior is to classify the user. The classification device according to any one of claims 1 to 5. The classification device according to any one of claims 1 to 6 , further comprising a providing unit that provides the content according to the classification result by the classification unit to the user. The classification device according to claim 7 , wherein the providing unit provides content selected according to an action performed by the user among the contents according to the classification result by the classification unit. It is a classification method executed by the classification device.
For each action of the reaching user who has performed a predetermined target action, an index value is calculated based on the value obtained by dividing the number of times the reaching user has performed the action by the number of times each user has performed the action. Calculation process and
Based on the index value calculated by the calculation step, each of the plurality of groups is divided into one of a plurality of groups in which the users who have performed the actions classified into each group are different in the possibility of performing the target action. classification method classifies an action, based on the classification result of the user to be classified is conducted action, characterized in that it comprises a classification step of classifying the user. On the computer
For each action of the reaching user who has performed a predetermined target action, an index value is calculated based on the value obtained by dividing the number of times the reaching user has performed the action by the number of times each user has performed the action. Calculation procedure and
Based on the index value calculated by the calculation procedure, each of the plurality of groups is divided into one of a plurality of groups in which the user who has performed the action classified into each group has a different possibility of performing the target action. classifying the behavior, based on the classification result of the user to be classified is conducted action classification program for executing a classification procedure to classify the user.

Images