JP2013235371A - Information processing apparatus, information processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing apparatus in which a rank having a large number of users is correctly extracted so as to present, to the users, a candidate for an effective new fee plan.SOLUTION: An information processing device: creates a randomized histogram from attribute information each of a plurality of users of a service and use history information of the service of the plurality of users; extracts a bin satisfying a prescribed condition as a user group; and extracts attribute information of the user group on the basis of attribute information of the users included in the user group. The information processing apparatus also calculates a group tendency value indicative of a tendency of use histories of users included in the user group as to a prescribed type of service, and a whole tendency value indicative of a tendency of use histories of the plurality of users as to the prescribed type. Thereafter, the information processing apparatus presents, when a difference between the group tendency value and the whole tendency value is equal to or more than a prescribed threshold value, a candidate for a new fee plan in which a unit price of a prescribed type is changed as to the user group.

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program for presenting a new charge plan candidate in a service.

  In urban areas, it is becoming difficult to own a car due to high annual maintenance costs such as parking fees. Moreover, since public transportation is developed in urban areas, automobiles are rarely used as daily means of transportation. However, there is certainly a demand to use a car only for a short time such as a weekend. In recent years, in response to such demand, car sharing services have been rapidly spread.

  The car sharing service is a service for sharing a vehicle among a plurality of members registered in advance, and has an advantage that an automobile can be used at a lower cost than a general rental car if used for a short time. The price plan of the car sharing service usually employs a time-based system with a shorter interval (for example, 15 minutes) than a rental car. In addition, there are companies that adopt a basic monthly fee and a distance-based system according to the distance traveled, and companies that set a unit-rate unit price for each grade of car model such as compact cars and sedans. is there.

  In the car sharing service, there is a case where a fee is discounted for a specific user group for the purpose of increasing the number of members. For example, it is possible to expect an increase in student membership by providing targeted student discounts to low income students.

  In this way, a new member can be acquired by setting a new price plan for a specific user group. Further, by narrowing down the user group, the premium feeling of the price plan can be increased and the attractiveness of the price plan can be improved. Furthermore, specifying the user group also has the effect of suppressing a decrease in sales when a large number of existing members move to a new rate plan.

  In order to set up such a new rate plan, it is necessary to first identify the user group in which the new rate plan is effective by analyzing the trend of the usage history.

  Here, for example, Patent Literature 1 discloses a method for extracting a user layer that matches a predetermined condition in retail marketing. In the method disclosed in Patent Document 1, two types of conditions are arranged in a matrix with a vertical axis and a horizontal axis, and information regarding the frequency that matches the conditions is displayed in each cell. Thereby, it is possible to easily set a better condition for extracting a user layer that matches a predetermined condition.

Japanese Patent No. 2905661

  In the method disclosed in Patent Document 1, when user information takes an integer value (real value) such as age or annual income, for example, the age is determined in advance such as “10-20” or “20-30”. Users are classified according to the given class. Therefore, for example, even when the number of users having the age of “15 to 25” is the largest, these users are distributed and classified into the classes of “10 to 20” and “20 to 30”. That is, there is a problem that it is impossible to correctly extract a class having a large number of users.

  In addition, the method disclosed in Patent Document 1 aims to specify an effective user layer in executing a campaign for a specific product. Therefore, in a rate plan that employs a pay-per-use system such as a car sharing service, no consideration is given to presenting what rate plan is effective for a specific user layer.

  An object of the present invention is to provide an information processing apparatus, an information processing method, and a program capable of correctly extracting classes having a large number of users and presenting effective new price plan candidates to those users. And

In order to achieve the above object, an information processing apparatus of the present invention is an information processing apparatus that presents a new price plan candidate in a service in which a price plan is determined at a predetermined type of unit price,
A user who creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group A group extractor;
A group attribute extraction unit that extracts attribute information of the user group based on the attribute information of the user included in the user group;
A group trend calculation unit that calculates a group trend value indicating a trend of the usage history of the service for the predetermined type of the user included in the user group based on the usage history information of the user included in the user group;
Based on the usage history information of the plurality of users, an overall trend calculation unit that calculates an overall tendency value indicating a trend of the usage history of the service for the predetermined type of the plurality of users;
When the difference between the group tendency value and the overall tendency value is equal to or greater than a predetermined threshold, a charge plan policy determination unit that presents a new charge plan candidate in which the predetermined unit price is changed for the user group; Have.

In order to achieve the above object, an information processing method of the present invention is an information processing method in an information processing apparatus that presents a new rate plan candidate in a service for which a rate plan is determined at a predetermined unit price. ,
Extraction that creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group Processing,
Based on the attribute information of the user included in the user group, a process of extracting the attribute information of the user group;
A process of calculating a group tendency value indicating a tendency of the use history of the service for the predetermined type of the user included in the user group based on the use history information of the user included in the user group;
Based on the usage history information of the plurality of users, a process of calculating an overall tendency value indicating a trend of the usage history of the service for the predetermined type of the plurality of users;
When the difference between the group tendency value and the overall tendency value is equal to or greater than a predetermined threshold value, a presentation process for presenting a new rate plan candidate with the predetermined unit price changed for the user group is included. .

In order to achieve the above object, the program of the present invention provides an information processing apparatus for presenting a new rate plan candidate in a service in which a rate plan is determined at a predetermined unit price.
Extraction that creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group Function and
A function of extracting attribute information of the user group based on the attribute information of the user included in the user group;
A function of calculating a group tendency value indicating a tendency of the use history of the service for the predetermined type of the user included in the user group based on the use history information of the user included in the user group;
A function for calculating an overall tendency value indicating a tendency of the usage history of the service for the predetermined type of the plurality of users based on the usage history information of the plurality of users;
When the difference between the group trend value and the overall trend value is equal to or greater than a predetermined threshold, a presentation function for presenting a new rate plan candidate in which the predetermined unit price is changed for the user group is realized. Let

  Since the present invention is configured as described above, it is possible to correctly extract classes having a large number of users and present effective new price plan candidates to those users.

It is a block diagram which shows the structure of one Embodiment of the information processing apparatus of this invention. 3 is a flowchart for explaining an operation of presenting a new price plan candidate in the information processing apparatus shown in FIG. 1. It is a block diagram which shows one structural example at the time of applying the information processing apparatus shown in FIG. 1 to a car sharing service system. It is a figure which shows an example of the information memorize | stored in the usage history memory | storage part shown in FIG. 3, (a) is a figure which shows an example of user attribute information, (b) is a figure which shows an example of usage history information.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of an embodiment of an information processing apparatus according to the present invention.

  As shown in FIG. 1, the information processing apparatus 10 according to the present invention includes a user group extraction unit 11, a group attribute extraction unit 12, a group tendency calculation unit 13, an overall trend calculation unit 14, a rate plan policy determination unit 15, It has. The usage history storage unit 20 stores information regarding each of the members who are a plurality of users who use the service. The service is, for example, a car sharing service.

  The usage history storage unit 20 stores user attribute information, which is member attribute information, and usage history information, which is member service usage history information. The usage history storage unit 20 stores at least one item of user attribute information such as sex, age, occupation, and address, for example. The usage history storage unit 20 stores a member ID (IDentification) that uniquely identifies each of a plurality of members and user attribute information in association with each other. In addition, the usage history storage unit 20 stores at least one item or more of information related to service usage history such as usage time, payment amount, and average moving speed as usage history information in association with the member ID. Note that the usage history information may be information related to one use, or may be statistical information in a past fixed period. The usage history information is updated every time the service is used or every certain period.

  The user group extraction unit 11 extracts a user group having a characteristic service usage tendency from the user attribute information and the usage history information stored in the usage history storage unit 20. Specifically, the user group extraction unit 11 creates a randomized histogram using random partitioning from the user attribute information and the usage history information stored in the usage history storage unit 20. Then, the user group extraction unit 11 extracts bins that satisfy a predetermined condition as user groups. The bins satisfying the predetermined condition are, for example, bins whose frequency exceeds a predetermined first threshold, bins whose frequency is the upper fixed number, and the like. Here, as random partitioning, user attribute information and usage history information are treated as multidimensional data, and a partitioning technique is used in which elements with close distances in a multidimensional space are aggregated in the same bin of the histogram with high probability. Use. As such a method, for example, <Shinji Kami et al., Theoretical Consideration of Residence Point Estimation by Spatial Division Using LSH, September 2011, IEICE Communication Society Conference Proceedings> (hereinafter non-patented) There is a spatial division used in the dwell point extraction method described in Document 1. However, since the method described in Non-Patent Document 1 is based on space division in the Euclidean space, it is necessary to appropriately perform the definition of the distance between the pieces of usage history information. For example, the weighting may be performed based on the division width of data expected in each item. Moreover, you may create a histogram using the other local sensitive hash (LSH: Locality Sensitive Hashing) using the randomization. In the user attribute information and the usage history information, there are items in which distances cannot be simply defined, such as occupation. Such items can be used as multidimensional data by defining the similarity between items and using the multidimensional scaling method.

  The group attribute extraction unit 12 extracts group attribute information that is a representative attribute of a member included in the user group extracted by the user group extraction unit 11. For extraction of group attribute information, a median value or an average value can be used for attribute information having a real number as a range. When it is desired to extract the group attribute information as a range, for example, the standard deviation σ is obtained, and (μ ± σ) or (μ ± 2σ) is extracted as the group attribute information with respect to the average value μ. When a plurality of items are used as group attribute information, the group attribute information is expressed as a set of the plurality of items.

  The group tendency calculation unit 13 calculates a group tendency value indicating the tendency of the service usage history of the user group based on the member usage history information included in the user group extracted by the user group extraction unit 11. Specifically, the group tendency calculation unit 13 obtains a group tendency value by obtaining a statistical value such as an average value or a median value of member usage history information included in the user group extracted by the user group extraction unit 11. calculate.

  Based on the usage history information stored in the usage history storage unit 20, the overall trend calculation unit 14 calculates an overall trend value indicating the trend of the usage history of all the members. Specifically, the overall trend calculation unit 14 calculates an overall trend value by obtaining a statistical value such as an average value or a median value from the usage history information of all the plurality of members. Note that the overall trend calculation unit 14 calculates an overall trend value at a timing when the information stored in the usage history storage unit 20 is updated or periodically.

  The rate plan policy determination unit 15 is a target for presenting a new rate plan candidate based on the group trend value calculated by the group trend calculation unit 13 and the overall trend value calculated by the overall trend calculation unit 14. The user group to be determined is determined, and new price plan candidates are presented. Specifically, the rate plan policy determination unit 15 determines whether to present a new rate plan candidate depending on whether the difference between the group trend value and the overall trend value is equal to or greater than a predetermined second threshold value. To do. When presenting a new rate plan candidate, if the service is used under the same conditions, the new rate plan candidate will be cheaper than the current rate plan for members included in the user group To do. The rate plan policy determination unit 15 includes a database 15a that stores a plurality of rate plans. In addition, the rate plan policy determination unit 15 recognizes the currently set rate plan among the plurality of rate plans stored in the database 15a. For example, when the rate plan is time-based pay-per-use, the rate plan policy determination unit 15 calculates the group trend value and the overall trend value based on the average value of service usage time. In this case, the plurality of data stored in the database 15a that the difference between the average usage time of the users included in the user group and the average usage time of all the members is equal to or greater than the second threshold. Among the price plans, a price plan whose unit price per hour is lower than that currently set is presented as a new price plan. The predetermined second threshold value may be a value set in advance by an operator or the like, or adaptively determined so that the number of new rate plan candidates to be finally presented falls within a certain range. May be. Then, the charge plan policy determination unit 15 outputs information indicating a new charge plan candidate to the outside. For example, the rate plan policy determination unit 15 includes a display screen (not shown), and displays group attribute information of the user group and information indicating a new rate plan candidate on the display screen.

  Below, the operation | movement which shows the candidate of a new rate plan in the information processing apparatus 10 comprised as mentioned above is demonstrated.

  FIG. 2 is a flowchart for explaining the operation of presenting a new price plan candidate in the information processing apparatus 10 shown in FIG.

  First, the user group extraction unit 11 creates a randomized histogram using random partitioning from user attribute information and usage history information stored in the usage history storage unit 20 (step S1).

  Then, the user group extraction unit 11 extracts bins that satisfy a predetermined condition as user groups (step S2).

  Next, the group attribute extraction unit 12 confirms whether there is an extracted user group (step S3).

  If there is no extracted user group as a result of the confirmation in step S3, the process is terminated.

  On the other hand, if there is an extracted user group as a result of the confirmation in step S3, the group attribute extraction unit 12 selects one of the unselected user groups from the extracted user groups (step S4).

  Next, the group attribute extraction unit 12 extracts group attribute information of the user group selected by the user group extraction unit 11 (step S5).

  Next, the group tendency calculation unit 13 calculates a group tendency value from member usage history information included in the user group selected by the user group extraction unit 11 (step S6).

  Next, the rate plan policy determination unit 15 determines that the difference between the group trend value calculated by the group trend calculation unit 13 and the overall trend value calculated by the overall trend calculation unit 14 is equal to or greater than a predetermined second threshold value. (Step S7).

  If the difference between the group trend value and the overall trend value is equal to or greater than the predetermined second threshold as a result of the confirmation in step S7, the charge plan policy determination unit 15 determines the group attribute information of the user group and the new charge plan. Information indicating the candidates is output (step S8).

  Next, the group attribute extraction unit 12 confirms whether there is an unselected user group among the extracted user groups (step S9).

  As a result of the confirmation in step S9, if there is no unselected user group among the extracted user groups, the process is terminated.

  On the other hand, if there is an unselected user group among the extracted user groups as a result of the confirmation in step S9, the process proceeds to step S4.

  Note that, as a result of the confirmation in step S7, when the difference between the group tendency value and the overall tendency value is smaller than the predetermined second threshold value, the process proceeds to step S9.

  Next, as a specific example, a case where the information processing apparatus of the present invention is applied to the presentation of a new rate plan candidate in the car sharing service system will be described.

  Here, in the car sharing service system, it is assumed that the charge plan P is determined by the basic charge B, and the hourly unit price Rt and the distance unit price Rd which are predetermined unit prices. Specifically, the rate plan P is calculated as (P = B + Rt × T + Rd × D) where the usage time is T and the travel distance is D.

  FIG. 3 is a block diagram showing a configuration example when the information processing apparatus 10 shown in FIG. 1 is applied to a car sharing service system. 3, the same components as those shown in FIG. 1 are denoted by the same reference numerals, and detailed description of the same components is omitted.

  The car sharing service system 100 performs various processes such as management of registered users, management of usage history, vehicle management, and billing processing in car sharing.

  As shown in FIG. 3, the usage history storage unit 20 is arranged in the car sharing service system 100, and stores user attribute information and usage history information of each of a plurality of members.

FIG. 4 is a diagram illustrating an example of information stored in the usage history storage unit 20 illustrated in FIG. 3, where (a) illustrates an example of user attribute information, and (b) illustrates an example of usage history information. FIG.
Here, in order to simplify the explanation, as shown in FIG. 4, the usage history storage unit 20 includes an age Y as user attribute information and an average moving speed S in the previous month as usage history information. Assume that it is stored in association with the ID.

  As described above, the user group extraction in the user group extraction unit 11 uses the space division method described in Non-Patent Document 1. At this time, since the unit is different between the age Y and the average moving speed S, the distance cannot be simply defined. For this reason, it is necessary to perform processing by converting the coordinates into appropriate coordinates in the Euclidean space. For example, when it is considered that space division with a width of about 10 years old and an average moving speed of about 20 (km / h) is preferable, a set of age and average moving speed (Y, S) is expressed as (Y, S). By converting to S / 2) and setting the width 10 as the division width, a spatial division suitable for the expected division width can be realized.

  In FIG. 4, the information corresponding to U1 and U3 is classified into different bins when a histogram is simply created, specifically when classified as 10 to 19 years old, 20 to 29 years old, for example. However, although the usage tendency is similar, it cannot be extracted as the same user group. However, the probability of U1 and U3 being classified into the same bin is increased by randomizing the space division and creating a plurality of histograms. When the histogram creation is completed, the user group extraction unit 11 extracts bins satisfying a predetermined condition from the bins as a user group. In the following description, it is assumed that U1 and U3 are classified into the same user group Gr1.

  Extraction of group attribute information by the group attribute extraction unit 12 for the user group Gr1 and calculation of a group tendency value by the group tendency calculation unit 13 are performed as follows. First, in the extraction of group attribute information, the average μ and standard deviation σ, which are user attribute information, are obtained. Since μ = 20 and σ = 2 in the user group Gr1, if μ ± 2σ is adopted as the group attribute, the ages 16-24 are extracted as group attribute information of the user group Gr1. Become. In addition, the trend of the usage history of the user group is performed by obtaining the average of items. Accordingly, here, the group tendency value of the user group Gr1 is 21 (km / h) ((18 + 24) / 2) on average between the average moving speed S of U1 and the average moving speed S of U3.

  As described above, the rate plan policy determination unit 15 compares the group tendency value with the overall tendency value. When the overall tendency value is calculated from the information shown in FIG. 4, the average moving speed is 31 (km / h) ((18 + 51 + 24) / 3). That is, the difference between the group tendency value and the overall tendency value is 10 (km / h). Here, assuming that the predetermined second threshold is 10 (km / h), the difference between the group tendency value of the user group of the user group Gr1 and the overall tendency value is equal to or greater than the predetermined second threshold. Therefore, the rate plan policy determination unit 15 presents a new rate plan candidate for the user group Gr1. At this time, the tendency of the usage history of the user group Gr1 is slower than the tendency of the usage history of all the members. For this reason, it is considered that there are many times when parking and spending time slowly. For such users, it is more attractive that the unit price per hour is cheaper, and it is considered that this is an effective rate plan. Therefore, the rate plan policy determination unit 15 presents a new rate plan for changing (decreasing) the hourly unit price Rt. In addition, in order to suppress a change in the charge plan of a user whose average moving speed is fast among a plurality of users included in the same user group, it is possible to simultaneously increase the distance unit price Rd in addition to the decrease in the hour unit price Rt. . In addition, when a certain percentage of the target user group chooses the best plan after introducing the proposed new price plan, the number of users required to reduce sales and increase profits is presented at the same time. May be.

  In the present embodiment, the usage history storage unit 20 has been described as being provided in a device or system other than the information processing apparatus 10, but the usage history storage unit 20 is stored in the information processing apparatus 10. You may make it prepare.

  As described above, in the present embodiment, classes having a large number of users can be correctly extracted, and effective new price plan candidates can be presented to those users.

  The constituent elements of the present invention are not limited to those separated as functions shown in the above-described block diagram, and may be separated as appropriate for each function.

  In the present invention, the processing in the information processing apparatus records a program for realizing the function on a recording medium readable by the information processing apparatus in addition to the processing realized by the dedicated hardware described above. The program recorded on the recording medium may be read by the information processing apparatus and executed. The recording medium readable by the information processing apparatus refers to a transfer medium such as a flexible disk, a magneto-optical disk, a DVD, and a CD, as well as an HDD incorporated in the information processing apparatus.

DESCRIPTION OF SYMBOLS 10 Information processing apparatus 11 User group extraction part 12 Group attribute extraction part 13 Group tendency calculation part 14 Overall trend calculation part 15 Rate plan policy determination part 15a Database 20 Usage history storage part 100 Car sharing service system

Claims (9)

An information processing apparatus for presenting a new price plan candidate in a service for which a price plan is determined at a predetermined unit price,
A user who creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group A group extractor;
A group attribute extraction unit that extracts attribute information of the user group based on the attribute information of the user included in the user group;
A group trend calculation unit that calculates a group trend value indicating a trend of the usage history of the service for the predetermined type of the user included in the user group based on the usage history information of the user included in the user group;
Based on the usage history information of the plurality of users, an overall trend calculation unit that calculates an overall tendency value indicating a trend of the usage history of the service for the predetermined type of the plurality of users;
When the difference between the group tendency value and the overall tendency value is equal to or greater than a predetermined threshold, a charge plan policy determination unit that presents a new charge plan candidate in which the predetermined unit price is changed for the user group; And an information processing apparatus. The information processing apparatus according to claim 1,
The user group extraction unit is an information processing apparatus that uses a local sensitive hash when creating the histogram. The information processing apparatus according to claim 1 or 2,
The rate plan policy determination unit reduces a price of the predetermined type, and a new rate plan that is cheaper than the current rate plan when the service is used under the same conditions for the user group. An information processing device that presents a price plan. An information processing method in an information processing apparatus for presenting a new price plan candidate in a service for which a price plan is determined at a predetermined unit price,
Extraction that creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group Processing,
Based on the attribute information of the user included in the user group, a process of extracting the attribute information of the user group;
A process of calculating a group tendency value indicating a tendency of the use history of the service for the predetermined type of the user included in the user group based on the use history information of the user included in the user group;
Based on the usage history information of the plurality of users, a process of calculating an overall tendency value indicating a trend of the usage history of the service for the predetermined type of the plurality of users;
When the difference between the group tendency value and the overall tendency value is equal to or greater than a predetermined threshold value, a presentation process for presenting a new rate plan candidate with the predetermined unit price changed for the user group is included. Information processing method. The information processing method according to claim 4,
In the extraction process, an information processing method using a local sensitive hash when creating the histogram. The information processing method according to claim 4 or 5,
In the presenting process, by reducing the unit price of the predetermined type, when the service is used for the user group under the same conditions, a charge plan that is cheaper than the current charge plan is set as a new charge plan. An information processing method that is a process to be presented. An information processing device that presents new price plan candidates in a service for which a price plan is determined at a predetermined unit price,
Extraction that creates a randomized histogram from attribute information of each of a plurality of users of the service and usage history information of each of the plurality of users and extracts bins satisfying a predetermined condition as a user group Function and
A function of extracting attribute information of the user group based on the attribute information of the user included in the user group;
A function of calculating a group tendency value indicating a tendency of the use history of the service for the predetermined type of the user included in the user group based on the use history information of the user included in the user group;
A function for calculating an overall tendency value indicating a tendency of the usage history of the service for the predetermined type of the plurality of users based on the usage history information of the plurality of users;
When the difference between the group trend value and the overall trend value is equal to or greater than a predetermined threshold, a presentation function for presenting a new rate plan candidate in which the predetermined unit price is changed for the user group is realized. Program to let you. The program according to claim 7,
In the extraction function, a program that uses a local sensitive hash when creating the histogram. In the program according to claim 7 or 8,
The presenting function reduces, as a new price plan, a price plan that is lower than the current price plan when the service is used under the same conditions for the user group by reducing the unit price of the predetermined type. A program that is a function to present.

Images