WO2017130488A1 - 情報処理装置、情報処理方法、およびプログラム - Google Patents
情報処理装置、情報処理方法、およびプログラム Download PDFInfo
- Publication number
- WO2017130488A1 WO2017130488A1 PCT/JP2016/081598 JP2016081598W WO2017130488A1 WO 2017130488 A1 WO2017130488 A1 WO 2017130488A1 JP 2016081598 W JP2016081598 W JP 2016081598W WO 2017130488 A1 WO2017130488 A1 WO 2017130488A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- measure
- customer
- contract
- state
- probability
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0637—Strategic management or analysis, e.g. setting a goal or target of an organisation; Planning actions based on goals; Analysis or evaluation of effectiveness of goals
- G06Q10/06375—Prediction of business process outcome or impact based on a proposed change
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
- G06Q30/015—Providing customer assistance, e.g. assisting a customer within a business location or via helpdesk
- G06Q30/016—After-sales
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0241—Advertisements
- G06Q30/0251—Targeted advertisements
- G06Q30/0254—Targeted advertisements based on statistics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/18—Legal services; Handling legal documents
Definitions
- the present disclosure relates to an information processing apparatus, an information processing method, and a program.
- the withdrawal prevention measure has been implemented even for customers with low effectiveness of the measure.
- the present disclosure proposes an information processing apparatus, an information processing method, and a program capable of presenting an index for successful measures for guiding a customer's continuation state to a target state.
- An information processing apparatus includes an output unit that outputs an indicator that a measure for leading to the target state is successful.
- the processor is based on past information that indicates a change in a customer's continuation state and past information that indicates whether or not a measure that leads the customer to the target state that has changed has been implemented. Proposed is an information processing method including controlling to output an index for the calculated measure that leads to the target state to be successful.
- a program is proposed for functioning as an output control unit that controls to output an indicator that the measure leading to the target state is successful.
- FIG. 1 is a diagram illustrating a schematic configuration of a system according to an embodiment of the present disclosure.
- FIG. It is a block diagram showing an internal configuration of a system by one embodiment of this indication. It is a block diagram showing an example of functional composition of a database and processing part of a server in one embodiment of this indication. It is a block diagram showing other examples of functional composition of a database of a server, and a treating part in one embodiment of this indication. It is a figure which shows the graphical model of the prediction model by this embodiment. It is a flowchart which shows the operation
- Non-Patent Document 1 describes a technique for predicting a customer's contract probability by solving a classification problem of whether or not a contract is concluded by machine learning. Such a technique can be applied to the above-described calculation of the departure probability.
- Non-Patent Document 1 “Probabilistic Modeling of a Sales Funnel to Prioritize Leads,” Brendan Duncan and Charles Elkan, Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Pages 1751-1758, 2015
- the withdrawal prevention measure has been implemented even for customers who have a high withdrawal probability but a low effectiveness of the withdrawal prevention measure.
- the above problem is solved by predicting, for each customer, the probability of success of the withdrawal prevention measure in addition to the customer's withdrawal probability.
- the marketer chooses a customer who has a high effect of the measure by presenting an indicator that the measure for guiding the customer's continuation state to the target state (also called a marketing measure) is successful. Is possible.
- the probability that it is such a customer is calculated and it is calculated.
- the probability of successful withdrawal prevention measures can be calculated in consideration.
- the system according to the present embodiment uses a probability model having the following characteristics to estimate the probability that each customer's state will change when a certain measure is implemented (determine the parameter). To do).
- the probability that the state change of each customer occurs when the measure is not implemented is represented by x i and the probability function of parameters, p (a ⁇ b
- p the probability that the state change of each customer is caused by the measure
- x i the probability that the state change of each customer is caused by the measure
- the system according to the present embodiment uses p (a ⁇ b
- the marketer can select a customer who implements the measure in descending order of the score corresponding to the probability calculated in this way, and can implement the measure for a customer who is likely to change the state due to the measure.
- customers can be selected in ascending order of the probability score calculated by the following equation 1.
- a i represents the current state of the customer
- a ′ represents a state that is not the purpose.
- the target state a target is assumed to shift to a new contract that increases profits, although the target state a target leaves the current state, for example, the contract contents are ranked up.
- the system according to the present embodiment generates a data entry for each customer at a certain point in time, “whether the contract has been withdrawn, whether the measure has been implemented, contract / customer data”, and learning of machine learning. Use as data.
- a vector obtained by vectorizing feature quantities is used as the data entry of the contract / customer data.
- x i is a feature vector of a contract / customer data of customer i.
- the probability of leaving without taking a leave prevention measure is p 1 (x i ), and the probability of being able to prevent leave by the leave prevention measure is p 2 (x i ).
- p 3 (x i ) be the probability that the intention to continue will not change, and each event will be expressed as shown in Table 1 below. Based on this expression, p 1 (x i ), p 2 (x i ), and p 3 (x i ) can be predicted from the learning data by maximum likelihood estimation.
- the following table 2 shows the customer configuration of each pattern in the presence or absence of a withdrawal prevention measure by telephone and the change in contract status.
- the annual flat-rate video distribution Web service is, for example, a contract system every year, and is a service that allows users to freely view videos provided by the service from a Web site for one year by paying an annual contract fee.
- a marketing measure to prevent withdrawal from the contract consider distributing gift coupons by mail.
- the coupon ID written in the coupon is entered on the designated screen of the Web service, the annual contract fee can be discounted. Since it takes a cost to create and mail a mailed gift with a coupon written therein, it is necessary to mail the coupon (hereinafter simply referred to as a coupon) only to customers who meet the cost.
- FIG. 1 is a diagram illustrating a schematic configuration of a system according to an embodiment of the present disclosure.
- a system 10 according to the present embodiment includes a client 100 and a server 300.
- the client 100 and the server 300 are connected by a network 200 and can communicate with each other.
- the client 100 can include, for example, a smartphone 100a, a personal computer 100b, and a tablet 100c.
- the client 100 is not limited to the illustrated example, and may include any type of terminal device having a function of inputting / outputting information to / from a user.
- the client 100 uses, for example, an image or sound in order to output information to the user.
- the client 100 may accept input of information from the user by an operation input of the terminal device, a voice indicating speech, or an image indicating a gesture or line of sight.
- the server 300 includes one or a plurality of server devices on the network.
- a plurality of server devices cooperate to realize the function of the server 300 described below, the whole of the plurality of server devices may be handled as a single information processing device.
- at least a part of the server device may be operated by an operator different from the operator of the server 300 described below.
- a part of the server 300 may be referred to as an external server that is not included in the system 10.
- at least a part of the server device has a database 310.
- the database 310 stores information related to customers and customer contract information.
- the network 200 includes various wired or wireless networks such as the Internet, a LAN (Local Area Network), or a mobile phone network.
- the network 200 may connect the client 100 and the server 300 and may connect a plurality of server devices included in the server 300.
- the network 200 may include a router or a hub that connects these networks to each other.
- FIG. 2 is a block diagram illustrating an internal configuration of a system according to an embodiment of the present disclosure.
- the client 100 may include a local storage 110, a communication unit 120, a processing unit 130, and an input / output unit 140.
- the server 300 can include a database 310, a communication unit 320, and a processing unit 330.
- each functional configuration will be further described.
- the terminal device functioning as the client 100 and one or more server devices included in the server 300 are realized by, for example, a hardware configuration of an information processing device described later.
- the local storage 110 is realized by, for example, a memory or a storage included in the terminal device.
- information provided from the server 300 via the network 200 and information input by the user via the input / output unit 140 are stored in the local storage 110 temporarily or continuously.
- the user can refer to the information provided from the server 300 or input a draft of the information provided to the server 300 even offline.
- the communication unit 120 communicates with the server 300 via the network 200.
- the communication unit 120 is realized by, for example, a communication device that performs communication on a network to which the client 100 is connected.
- the processing unit 130 is realized by a processor such as a CPU (Central Processing Unit) included in the terminal device, for example.
- the processing unit 130 executes processing for requesting information from the server 300 via the communication unit 120 based on information input by the user via the input / output unit 140. Further, for example, the processing unit 130 executes a process of outputting information to the user via the input / output unit 140 based on information provided from the server 300 via the communication unit 120. At this time, the processing unit 130 may execute a process of converting the provided information into an appropriate format according to the type of the input / output unit 140.
- a processor such as a CPU (Central Processing Unit) included in the terminal device, for example.
- the processing unit 130 executes processing for requesting information from the server 300 via the communication unit 120 based on information input by the user via the input / output unit 140. Further, for example, the processing unit 130 executes a process of outputting information to the user via the input / output unit 140 based on information provided from the server 300 via the communication unit 120
- the input / output unit 140 is realized by, for example, an input device such as a touch panel, a mouse, a keyboard, a microphone, or a camera (imaging device) included in the terminal device and an output device such as a display or a speaker.
- the input / output unit 140 may include only one of the input device and the output device. For example, information received from the server 300 via the communication unit 120 is displayed on the display included in the input / output unit 140 through processing by the processing unit 130. Further, for example, a user operation input acquired by a touch panel included in the input / output unit 140 is transmitted to the server 300 via the communication unit 120 through processing by the processing unit 130.
- the functions of the processing unit 130 and the input / output unit 140 as described above are the same as the functions of the processing unit and the input / output unit in a general terminal device, for example, they will be described in detail in the following description of the present embodiment. It may not be done. However, even in such a case, for example, if the information received from the server 300 is characteristic, the function of the processing unit 130 or the input / output unit 140 in the client 100 is also in that the information is processed and output. It can be characteristic compared to their functions in a general terminal device.
- the database 310 is realized by, for example, a memory or a storage included in the server device. As described above, the database 310 stores information related to customers and customer contracts. More specific types of information stored in the database 310 may vary depending on the content of the service provided by the server 300.
- the communication unit 320 communicates with the client 100 via the network 200. Further, the communication unit 320 may communicate with an external server via the network 200.
- the communication unit 320 is realized by a communication device that performs communication on a network to which the server 300 is connected, for example.
- the processing unit 330 is realized by a processor such as a CPU included in the server device, for example.
- the processing unit 330 acquires information from the database 310 based on information received from the client 100 via the communication unit 320, processes the acquired information as necessary, and then performs communication with the communication unit 320. The process to transmit to the client 100 is executed.
- the functional configuration of the server 300 described above can be realized by being distributed to the plurality of server devices.
- the function of the database 310 may be realized intensively by any server device, or may be realized by operating a database distributed to a plurality of server devices in an integrated manner.
- the function of the processing unit 330 may be realized by operating processors distributed to a plurality of server devices in an integrated manner.
- the functions of the processing unit 330 described below can be realized by being distributed in series or in parallel to a plurality of server devices regardless of the functional block categories defined for the description.
- FIG. 3 is a block diagram illustrating a functional configuration example of a database and a processing unit of a server according to an embodiment of the present disclosure.
- customer / contract data 3101 and learning data 3102 are shown as functions of the database 310 of the server 300
- a learning unit 3301, a prediction unit 3302, and a calculation unit 3303 are shown as functions of the processing unit 330. It is shown.
- a user I / F unit 1401 is shown as a function of the input / output unit 140 of the client 100.
- FIG. 4 is a block diagram illustrating another example of the functional configuration of the database and the processing unit of the server according to an embodiment of the present disclosure.
- a calculation unit 1301 is shown as a function of the processing unit 130 of the client 100.
- the calculation unit is on the client side is also assumed.
- the customer / contract data 3101 functions as master data of a customer / contract handled by a service provided by the server 300.
- the customer / contract data may include, for example, customer information, contract information, contract status information, measure implementation information, and the like.
- the customer information is information representing customer characteristics, which includes customer attributes and an action log (video viewing, etc.) on the customer's Web service. More specifically, for example, the customer information may include age, sex, address, family composition, previously viewed videos, genres of frequently viewed videos, and the like.
- the learning unit 3301 generates a feature vector of the customer based on the customer information and the contract information.
- the contract information is information related to the contract and is information representing the characteristics of the contract.
- the contract information may include the contract continuation year, contract date, past withdrawal, payment method, contract agent, discount usage, and the like.
- the contract status information is information indicating a contract status such as whether or not a contract is in progress, and is used as a customer status teacher label in the learning unit 3301.
- the contract status information may include whether or not the contract is currently being contracted, and the date and time when the contract is terminated.
- the measure execution information is information representing a measure execution history such as whether or not a measure for preventing withdrawal has been executed in the past, and is treated as input information different from the feature vector.
- the measure implementation information may include whether or not a coupon has been mailed, and when mailed, the date and time of shipment.
- the customer / contract data 3101 data related to the contract as described above is registered in association with an ID unique to each customer, for example.
- the learning data 3102 includes feature quantities of customer / contract data registered in the customer / contract data 3101 (hereinafter also referred to as customer / contract feature quantities).
- the customer / contract data 3101 is generated by the learning unit 3301 using, for example, the customer / contract data 3101.
- the learning data 3102 generation process is executed once a month, for example.
- the customer / contract feature quantity may be a vector extracted from each data item for certain customer / contract data (specified by the customer ID).
- a feature vector can be stored in association with the customer ID.
- One feature quantity is basically stored for one customer. Accordingly, the learning data 3102 can be used as information representing the current state of the customer, for example. Details of the processing of the learning unit 3301 for generating the feature amount will be described later.
- the learning unit 3301 performs processing for calculating (generating) various parameters of the prediction model by machine learning using past customer / contract data as learning data.
- the learning unit first generates learning data 3102 from the customer / contract data 3101, and then generates a prediction model parameter based on the learning data 3102. These processes are periodically executed, for example, once a month, and the parameters held in the prediction unit 3302 are updated with the calculated parameters.
- the learning data 3102 includes, for example, the past Generated for customers who renewed or left the contract in 11 months. In other words, customers that will be subject to renewal within the next month are not included.
- the learning data 3102 is composed of a set of customer data entries generated for each target customer.
- the customer data entry includes three elements, “whether or not contracted”, “whether or not a measure has been implemented”, and “feature vector”.
- Whether the contract is in progress can be created from the contract status information included in the customer / contract data 3101. Further, whether or not a measure has been implemented in the past 11 months can be created from the measure implementation information included in the customer / contract data 3101.
- the feature vector is a vector extracted from the customer / contract data 3101 and is an example of feature data. Such a feature vector may be generated, for example, by simply connecting data items.
- a feature vector is generated for each customer ID that uniquely identifies the customer / contract data 3101, and a combination such as “customer ID, whether contract is in progress, whether measures have been implemented, feature vector” May be stored as learning data 3102.
- Learning unit 3301 first converts each item (age, address, etc.) of customer information and contract information into a vector called a one-of-k vector.
- the one-of-k vector is a k-dimensional vector, in which only one element takes a value of 1 and the remaining (k-1) elements take a value of 0.
- the learning unit 3301 enumerates possible values of the item, creates a vector having the same dimension as the number of possible values, and determines a dimension corresponding to the possible values. In vectorization, the dimension corresponding to the item value is set to 1, and the rest is set to 0, thereby converting the item value to a one-of-k vector.
- 82-dimensional vectors from 18 to 99 years are prepared as possible values, and the age from 18 years is made to correspond in order from the first dimension.
- an 82-dimensional one-of-k vector is created in which the third dimension is 1 and the rest is 0.
- a one-of-k vector is created for each item of customer information and contract information, and a vector concatenated in a fixed order can be used as a customer feature vector.
- action history information such as videos that have been viewed in the past, prepare a vector of the dimension of the type of video, create a vector with the dimension corresponding to the video viewed in the past as 1, and the remaining dimensions as 0 May be.
- the learning unit 3301 functions as a generation unit that performs machine learning using the learning data 3102 and calculates (generates) various parameters. For example, the learning unit 3301 calculates various parameters used for the probability prediction model used in the measure success probability prediction process by the prediction unit 3302 by machine learning.
- learning of the probability prediction model by the learning unit 3301 will be specifically described. Note that the learning method described below is an example and is not necessarily limited thereto.
- the learning unit 3301 first defines the following variables.
- i is a customer ID.
- the learning unit 3301 determines the probability of each event as follows.
- the graphical model 20 of the following prediction model is shown in FIG.
- f h (x; w h ), f y1 (x; w y1 ), and f y2 (x; w y2 ) are parametric functions that take values of 0 or more and 1 or less, for example, logistic Regression models and neural networks can be used.
- the learning unit 3301 determines (calculates) various parameters (w h , w y1 , w y2 ) by maximum likelihood estimation.
- the maximum likelihood estimation is performed, for example, by executing a gradient descent method.
- An example of likelihood estimation is shown in Equation 2 below.
- the prediction unit 3306 uses the parameters learned by the learning unit 3301 to predict “continuation probability when measure is implemented—continuation probability when measure is not implemented” by the prediction model.
- the prediction unit 3306 uses the customer / contract data 3101 in the same processing as the learning unit 3301 to generate a feature vector for a customer who will reach the contract renewal time within the next month.
- the prediction unit 3306 uses the parameters learned by the learning unit 3301 to calculate “probability of leaving when no measure is taken” by f h (x; w h ), and f y1 (x i ; w
- the “probability that the withdrawal can be prevented when the measure is taken” is calculated by y1 ), and the “probability that the measure is continued even if the measure is taken” is calculated by f y2 (x i ; w y2 ).
- the calculated predicted value is output to the calculation unit 3303.
- Non-Patent Document 2 “A Contextual-Bandit Approach to Personalized News Article Recommendation, ”Lihong Li, Wei Chu, John Langford, and Robert E. Schapire, Proceedings of the 19th International Conference on World Wide Web, Pages 661 ⁇ 670, 2010
- the calculation unit 3303 calculates the numerical value requested from the user I / F unit 1401 of the client 100 using the probability output by the prediction unit 3302. In addition, the calculation unit 3303 generates data to be output to the user I / F unit 1401 such as a display included in the input / output unit 140 of the client 100 and transmits the data to the client 100 from the communication unit 320. Such data is, for example, a result image including calculated numerical values.
- the information output method in the client 100 is not limited to image display. For example, audio output may be used together with or instead of image display.
- the calculation unit 3303 calculates the profit rate for each customer based on the measure success probability, measure cost, and customer profit (subtract the discount amount as necessary), and arranges the calculated values as scores in descending order.
- a list image is generated and transmitted to the client 100.
- the measure success probability can be calculated by using the probability calculated by the prediction unit 3302 by “continuation probability when measure is executed ⁇ continuation probability when measure is not executed”.
- the user I / F unit 1401 of the client 100 accepts input from the user, transmits input data from the communication unit 120 to the server 300, and presents information transmitted from the server 300 to the user.
- FIG. 6 is a flowchart showing an operation process of the learning unit 3301 according to this embodiment.
- the learning unit 3301 selects a customer who has renewed or left a contract in the past 11 months based on the customer / contract data (step S103), and the customer is selected for the selected customer.
- a data entry is created, and the set is used as learning data 3102 (step S106).
- the customer data entry includes three elements, “whether or not contracted”, “whether or not a measure has been implemented”, and “feature vector”.
- feature vector creation process will be described later with reference to FIG.
- the learning unit 3301 generates a prediction model parameter from the learning data 3102 (step S109), and updates the parameter of the prediction unit 3302 with the generated parameter (step S112).
- FIG. 7 is a flowchart showing a feature vector generation process by the learning unit 3301. As shown in FIG. 7, the learning unit 3301 creates a one-of-k vector for each item of customer information and contract information (step S123). One-of-k vector generation processing will be described later with reference to FIG.
- the learning unit 3301 connects the one-of-k vectors of the items in a predetermined order to create a feature vector (step S126).
- FIG. 8 is a flowchart showing a one-of-k vector creation process according to this embodiment.
- the learning unit 3301 acquires information indicating the value X of the item A from the customer information and the contract information (Step S133).
- the learning unit 3301 refers to the dictionary data in which natural numbers are assigned in order from 1 to the values that can be taken by the item A, and acquires the natural number i assigned to X (step S136). For example, when converting age into a one-of-k vector, 82-dimensional vectors from 18 to 99 are prepared in advance as possible values, and the learning unit 3301 assigns the value “20 years” to the item “age” Get the natural number 3 given.
- the learning unit 3301 creates a k-dimensional vector in which the i-th dimension is 1 and the other dimensions are 0 (step S139). That is, for example, when converting an item whose age is 20 years, the learning unit 3301 creates an 82-dimensional one-of-k vector in which the third dimension is 1 and the remaining is 0.
- FIG. 9 is a flowchart showing the operation processing of the prediction unit 3302 according to this embodiment. As shown in FIG. 9, first, the prediction unit 3302 selects a customer who will reach the contract renewal time within the next month (step S143).
- the predicting unit 3302 creates a feature vector for the selected customer in the same manner as the processing by the learning unit 3301 (step S146).
- the prediction unit 3302 inputs a feature quantity vector to the prediction model using the parameter generated by the learning unit 3301, and calculates each probability (step S149).
- the prediction part 3302 transmits each probability calculated for every customer to the user I / F part 1401 contained in the input / output part 140 of the client 100 via the communication part 320 (step S152).
- Example of information presentation screen >> Subsequently, an example of information presented in an embodiment of the present disclosure will be described with reference to an example of a screen displayed on a display (user I / F unit 1401) included in the input / output unit 140 in the client 100, for example. To do. In the following explanation, examples of information presented when selecting target customers to implement withdrawal prevention marketing measures for the flat-rate video distribution service are explained. Information can be similarly presented in selecting a customer.
- the calculation unit 3303 of the server 300 calculates the profit rate based on the measure success probability, the measure cost, and the customer benefit, and the customer list screen in which the customer IDs are arranged in descending order of the profit rate value. It is displayed on the user I / F unit 1401 of the client 100 and can assist the marketing staff (user) to select the target customer for the measure.
- FIG. 10 is a diagram showing an example of a customer list screen displayed in the present embodiment.
- a customer list 1101, a download number designation field 1102, a download button 1103, an information input field 1104, and a calculation result 1105 are displayed on the screen 1100.
- the customer list 1101 is “continuation probability when the measure is implemented ⁇ continuation probability when the measure is not implemented” (substantial measure success probability). ) For each customer, the customer IDs are arranged in descending order of the score. The marketer refers to the customer list 1101 and selects a customer who implements the measure.
- the marketer selects a customer who has a high measure success probability (that is, a probability of prevention of withdrawal by the measure), that is, selects a predetermined upper number of people.
- the marketer transmits the information about the selected customer to the measure executor, for example, when the measure is “telephone”, to the call center.
- a download number designation field 1102 and a download button 1103 are provided on the screen 1100.
- the marketer inputs the target customers to be implemented to be communicated to the call center into the number-of-downloads designation field 1102 in the form of the top number of people, and selects the download button 1103.
- the client 100 transmits the input information to the server 300 via the network 200.
- the server 300 transmits customer IDs (with detailed information such as telephone numbers and addresses, if necessary) to the call center (not shown) as information indicating the person to be implemented according to the instructions of the marketer. .
- the calculation unit 3303 calculates the number of people considering the profit ratio as reference information when the marketer considers how many people in the top should be selected as the target of the measure implementation. It is also possible to present it.
- the marketer inputs the default price of the customer profit, the cost per measure, the discount amount of the coupon, and the target profit rate, and information on how many people should implement the measure. Obtainable.
- the value input in the information input field 1104 is transmitted to the server 300, and the calculation unit 3303 calculates the following formula 3.
- the calculation unit 3303 calculates the profit amount for each customer by the above formula 3, and calculates the profit rate by adding them in order from the top. Then, the calculation unit 3303 calculates how many people satisfy the designated target profit rate, and transmits the calculation result to the client 100.
- the client 100 displays the received calculation result 1105 on the screen 1100. As a result, the marketer knows how many people in the top level will implement the measure to meet the target profit rate, and can be used as a reference when selecting the target customers for the measure.
- the prediction certainty factor may be displayed side by side along with the predicted measure success probability score in the customer list.
- FIG. 11 shows an example of a screen on which the prediction certainty factors are also arranged.
- a customer list 1201 a download number designation column 1202, a download button 1203, an information input column 1204, and a calculation result 1205 are displayed on the screen 1200.
- customer IDs, scores, and certainty levels are displayed in association with each other.
- FIG. 12 is a diagram illustrating another example of the customer list screen displayed in the present embodiment.
- a customer list 1301 and a profit forecast graph 1302 are displayed on the screen 1300.
- customer IDs, scores, future customer profits, and profit estimates are displayed in association with each other.
- the customer list 1301 is arranged in the descending order of future customer profits.
- Customer profit is the product of the estimated value of the contract continuation year and the profit per year.
- the predicted value of the contract continuation year may be calculated by the machine learning by the calculation unit 3303 as a regression problem in which the feature quantity vector excluding the contract continuation year of the contract information is input and the output is the contract continuation year. Is possible.
- the profit forecast graph 1302 displays the expected profit rate and the expected profit as a graph when the measure is implemented from a customer with a higher score.
- the expected profit rate can be calculated by, for example, the following formula 4.
- the horizontal axis of the profit forecast graph 1302 is the number of customers targeted for the measure (number of people from the top score).
- the profit expectation graph 1302 shown on the screen 1300 can be presented after information necessary for calculation such as a coupon discount is input by the marketer.
- the screen 1300 does not display the download number designation field 1102 and the download button 1103 as shown in the screen 1100, but may include a screen including these.
- FIG. 13 shows an example of a customer information detail screen displayed when a customer ID included in the customer list 1301 on the screen 1300 is selected.
- the screen 1400 displays a customer ID, a probability of withdrawal, future customer profits, information for each measure, and customer details (address, age, gender, duration of contract, etc.). Yes.
- each measure shows the probability of withdrawal prevention, measure costs, expected profits, and expected effects, thereby visualizing the effects for each measure.
- Each information for each measure is obtained by the calculation unit 3303 as follows.
- Continuation probability when measures are implemented-Continuity probability when measures are not implemented-Cost Cost per measure-Expected profit: ⁇ (Predicted value of customer profit)-(Coupon discount) ⁇ ⁇ ((Continuation probability when measures are implemented – Continuity probability when measures are not implemented) -Expected effect: ⁇ (Predicted value of customer profit)-(Coupon discount) ⁇ x (Continuation probability when measures are implemented-Continuity probability when measures are not implemented)-(Cost per measure)
- the calculation unit 3303 calculates f y1 (x i ; w y1 ) and f y2 learned for each type of measure. Using (x i ; w y2 ), calculate “prohibition of withdrawal” for each type of measure.
- FIG. 14 shows a formula for calculating the likelihood of such a model.
- the visualization of the effect for each measure is not limited to the example shown in FIG. 13.
- a coupon having the maximum expected effect is selected for each customer and presented to the user when the expected effect is larger than the threshold value. May be.
- an “item, value” pair that works positively on the probability of withdrawal prevention and an “item, value” work that works negatively A pair may be output.
- a description will be given with reference to FIG.
- FIG. 15 is a diagram illustrating an example of a screen that presents items that affect the probability of leaving prevention according to the present embodiment.
- the screen 1500 displays a pair of “item, value” that works positively on the probability of preventing the withdrawal of a coupon measure (that is, the probability that withdrawal can be prevented when a measure is taken) and a “item, The “value” pair is displayed.
- the marketer knows that there is an effect in enhancing the drama for women and in appealing the drama that starts after the contract renewal period.
- FIG. 16 is a block diagram illustrating a hardware configuration example of the information processing apparatus according to the embodiment of the present disclosure.
- the illustrated information processing apparatus 900 can realize, for example, the server 300 and the client 100 in the above-described embodiment.
- the information processing apparatus 900 includes a CPU (Central Processing unit) 901, a ROM (Read Only Memory) 903, and a RAM (Random Access Memory) 905.
- the information processing apparatus 900 may include a host bus 907, a bridge 909, an external bus 911, an interface 913, an input device 915, an output device 917, a storage device 919, a drive 921, a connection port 923, and a communication device 925.
- the information processing apparatus 900 may include a processing circuit such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array) instead of or in addition to the CPU 901.
- DSP Digital Signal Processor
- ASIC Application Specific Integrated Circuit
- FPGA Field-Programmable Gate Array
- the CPU 901 functions as an arithmetic processing device and a control device, and controls all or a part of the operation in the information processing device 900 according to various programs recorded in the ROM 903, the RAM 905, the storage device 919, or the removable recording medium 927.
- the ROM 903 stores programs and calculation parameters used by the CPU 901.
- the RAM 905 primarily stores programs used in the execution of the CPU 901, parameters that change as appropriate during the execution, and the like.
- the CPU 901, the ROM 903, and the RAM 905 are connected to each other by a host bus 907 configured by an internal bus such as a CPU bus. Further, the host bus 907 is connected to an external bus 911 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 909.
- PCI Peripheral Component Interconnect / Interface
- the input device 915 is a device operated by the user, such as a mouse, a keyboard, a touch panel, a button, a switch, and a lever.
- the input device 915 may be, for example, a remote control device that uses infrared rays or other radio waves, or may be an external connection device 929 such as a mobile phone that supports the operation of the information processing device 900.
- the input device 915 includes an input control circuit that generates an input signal based on information input by the user and outputs the input signal to the CPU 901. The user operates the input device 915 to input various data and instruct processing operations to the information processing device 900.
- the output device 917 is configured by a device capable of notifying the acquired information to the user using a sense such as vision, hearing, or touch.
- the output device 917 can be, for example, a display device such as an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display, an audio output device such as a speaker or headphones, or a vibrator.
- the output device 917 outputs the result obtained by the processing of the information processing device 900 as video such as text or image, sound such as sound or sound, or vibration.
- the storage device 919 is a data storage device configured as an example of a storage unit of the information processing device 900.
- the storage device 919 includes, for example, a magnetic storage device such as an HDD (Hard Disk Drive), a semiconductor storage device, an optical storage device, or a magneto-optical storage device.
- the storage device 919 stores, for example, programs executed by the CPU 901 and various data, and various data acquired from the outside.
- the drive 921 is a reader / writer for a removable recording medium 927 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and is built in or externally attached to the information processing apparatus 900.
- the drive 921 reads information recorded on the attached removable recording medium 927 and outputs the information to the RAM 905.
- the drive 921 writes a record in the attached removable recording medium 927.
- the connection port 923 is a port for connecting a device to the information processing apparatus 900.
- the connection port 923 can be, for example, a USB (Universal Serial Bus) port, an IEEE 1394 port, a SCSI (Small Computer System Interface) port, or the like.
- the connection port 923 may be an RS-232C port, an optical audio terminal, an HDMI (registered trademark) (High-Definition Multimedia Interface) port, or the like.
- the communication device 925 is a communication interface configured with, for example, a communication device for connecting to the communication network 931.
- the communication device 925 can be, for example, a communication card for LAN (Local Area Network), Bluetooth (registered trademark), Wi-Fi, or WUSB (Wireless USB).
- the communication device 925 may be a router for optical communication, a router for ADSL (Asymmetric Digital Subscriber Line), or a modem for various communication.
- the communication device 925 transmits and receives signals and the like using a predetermined protocol such as TCP / IP with the Internet and other communication devices, for example.
- the communication network 931 connected to the communication device 925 is a network connected by wire or wireless, and may include, for example, the Internet, a home LAN, infrared communication, radio wave communication, satellite communication, or the like.
- Each component described above may be configured using a general-purpose member, or may be configured by hardware specialized for the function of each component. Such a configuration can be appropriately changed according to the technical level at the time of implementation.
- Embodiments of the present disclosure include, for example, an information processing device (server or client) as described above, a system, an information processing method executed by the information processing device or system, a program for causing the information processing device to function, and It may include a non-transitory tangible medium on which the program is recorded.
- a withdrawal prevention marketing measure for maintaining a continuous contract is used as an example of a marketing measure that promotes a change in the state of the customer, but this embodiment is not limited to this.
- the system according to the present embodiment can be applied to the following cases.
- a customer with a low “continuity probability when measures are implemented ⁇ continuity probability when measures are not implemented” is displayed to the marketer, and the marketer uses the current withdrawal prevention measures. It can also be used as a clue to think about measures to prevent withdrawal that is effective for customers who are not effective.
- the user I / F unit 1401 of the client 100 implements the “probability when the measure is implemented—the probability of continuation when the measure is not implemented” as a customer with a low “continuation probability when the measure is not implemented”.
- a certain number may be displayed in ascending order of “continuation probability when not doing”. In this case, it may be to create a customer list so that the variation of x i so that it does not contain a similar customer become rich.
- the characteristics of customers with low “continuation probability when measures are implemented—continuation probability when measures are not implemented” are commonly used for contracts and customer data that tend to appear in common. It may be presented as a set of elements. At this time, it may be possible to consider whether or not there is a large difference between the characteristics of the customer who has a high “continuation probability when the measure is implemented ⁇ continuation probability when the measure is not implemented”.
- the certainty of prediction can be calculated by any one of the following (1) to (3).
- (1) The average of the number of customers with a certain degree of similarity or a certain number of similar top ranks in the learning data (in this case, the higher the number, the higher the certainty of prediction)
- (2) The magnitude of variation in withdrawal or continuation among similar customers in the learning data (in this case, the smaller the variation, the higher the certainty of prediction)
- (3) Relationship between the posterior probability of the parameter of p 2 (x i ) and x i in this case, the higher the Bayesian posterior probability of the parameter used for estimating x i , the higher the certainty of prediction
- the certainty is high from the viewpoint (1) above but the certainty is low from the viewpoint (2), it is considered difficult to predict with the current contract / customer data.
- the marketer may be encouraged to acquire new features to find.
- this technique can also take the following structures.
- (1) To the target state calculated based on past information indicating a change in the customer's continuation state and past information indicating whether or not a measure that leads the customer to the target state has been implemented.
- An information processing apparatus comprising an output unit that outputs an indicator that a measure for guiding is successful.
- (2) The information processing apparatus according to (1), wherein the output unit calculates and outputs a measure success probability as an index that the measure for leading to the target state is successful for each customer currently in a continuous state.
- the continuation state of the customer is a contract state
- the change in continuation state is withdrawal of a contract
- the information processing apparatus according to (2), wherein the target state is a state in which the contract state is maintained.
- the information processing apparatus includes: A learning unit that generates a parameter of a prediction model based on past information related to the state, past information related to the implementation of the measure, and a feature amount related to the customer or the continuation state of the customer; Using the generated parameters, a prediction unit that calculates the probability of contract withdrawal when the measure is not implemented, and the probability of contract withdrawal prevention when the measure is implemented, in the customer who is currently in a continuous state; With The output unit calculates the product of the probability of contract withdrawal when the measure is not implemented and the probability of contract withdrawal prevention when the measure is implemented for each customer who is currently in a continuous state as an indicator that the measure is successful.
- the information processing apparatus according to (3).
- the prediction unit calculates a certainty factor of prediction, The information processing apparatus according to (6), wherein the output unit outputs the score together with a certainty of prediction.
- the output unit calculates a profit rate at the time of implementing the measure based on the revenue forecast at the time of contract maintenance and the cost of the measure, and outputs the profit rate to the user.
- the information processing apparatus described.
- the output unit outputs a prediction graph of the profit rate to a user.
- the output unit presents to the user a feature that works positively and a feature that works negatively with respect to the probability of preventing contract withdrawal when the measure is implemented.
- the continuation state of the customer is a contract state
- the change in the continuation state is a contract upgrade
- Processor To the target state calculated based on past information indicating a change in the customer's continuation state and past information indicating whether or not a measure that leads the customer to the target state has been implemented. An information processing method including controlling to output an indicator that a guiding measure is successful.
- (13) Computer To the target state calculated based on past information indicating a change in the customer's continuation state and past information indicating whether or not a measure that leads the customer to the target state has been implemented.
Abstract
Description
1.本開示の一実施形態によるシステムの概要
1-1.背景
1-2.アルゴリズムの概要
1-3.システム構成
(1-3-1.クライアントの構成)
(1-3-2.サーバの構成)
2.機能構成
2-1.データベースの構成例
2-2.処理部の構成例
3.動作処理
3-1.学習部3301の動作処理
3-2.予測部3302の動作処理
4.情報提示画面例
5.ハードウェア構成
6.まとめ
<1-1.背景>
近年では、離脱可能性の高い顧客の選択について、契約・顧客データに基づいて機械学習によって離脱確率を予測し、予測した離脱確率の高い順に離脱防止施策を実施することが行われる。この際、契約・顧客データを入力とし、離脱するか否かの分類問題として機械学習を適用する。例えば、下記非特許文献1では、契約が成約するか否かの分類問題を機械学習的に解くことにより、顧客の成約確率を予測する技術について述べられている。かかる技術は、上述の離脱確率の算出に適用することができる。
非特許文献1:”Probabilistic Modeling of a Sales Funnel to Prioritize Leads,” Brendan Duncan
and Charles Elkan, Proceedings of the 21th ACM SIGKDD International Conference
on Knowledge Discovery and Data Mining, Pages 1751-1758, 2015
まず、顧客の継続状態を目的の状態へ導くための施策(マーケティング施策とも称す)が成功する指標の算出アルゴリズムについて説明する。例えば、本実施形態によるシステムは、ある時点(時点t)で顧客毎に下記データエントリを生成し、その集合を機械学習の学習データとして使用する。
図1は、本開示の一実施形態によるシステムの概略的な構成を示す図である。図1を参照すると、本実施形態に係るシステム10は、クライアント100と、サーバ300とを含む。クライアント100とサーバ300とは、ネットワーク200によって接続され、相互に通信が可能である。
ローカルストレージ110は、例えば端末装置が有するメモリまたはストレージによって実現される。ローカルストレージ110には、例えば、ネットワーク200を介してサーバ300から提供された情報や、入出力部140を介してユーザによって入力された情報が、一時的または持続的に格納される。ローカルストレージ110に格納された情報を利用して、ユーザは、オフラインでもサーバ300から提供された情報を参照したり、サーバ300に提供する情報の下書きを入力したりすることができる。
データベース310は、例えばサーバ装置が有するメモリまたはストレージによって実現される。上記の通り、データベース310には、顧客や顧客の契約に関する情報が格納される。データベース310に格納される情報のより具体的な種類は、サーバ300によって提供されるサービスの内容によって異なりうる。
続いて、サーバ300のデータベース310および処理部330の機能構成について図3およびを参照して説明する。
(顧客・契約データ3101)
顧客・契約データ3101は、サーバ300によって提供されるサービスで扱われる顧客・契約のマスタデータとして機能する。顧客・契約データは、例えば、顧客情報、契約情報、契約状態情報、施策実施情報などを含みうる。顧客情報は、顧客属性および顧客のWebサービス上での行動ログ(動画視聴等)から成る、顧客の特徴を表す情報である。より具体的には、例えば、顧客情報は、年齢、性別、住所、家族構成、過去に視聴した動画、よく見る動画のジャンルなどを含みうる。詳細は後述するが、学習部3301では、顧客情報と契約情報に基づいて顧客の特徴量ベクトルを生成する。
学習データ3102は、顧客・契約データ3101に登録された顧客・契約データの特徴量(以下、顧客・契約特徴量とも称す)を含む。顧客・契約データ3101は、学習部3301により、例えば顧客・契約データ3101を用いて生成される。学習データ3102の生成処理は、例えば1カ月に1度実行される。具体的には、例えば顧客・契約特徴量は、ある顧客・契約データ(顧客IDによって特定される)について、それぞれのデータの項目から抽出されるベクトルでありうる。学習データ3102では、このような特徴量のベクトルが、顧客IDに関連付けて格納されうる。特徴量は、基本的に一人の顧客に対して1つ格納される。従って、学習データ3102は、例えば顧客の現在の状態を表す情報として利用されうる。なお、特徴量を生成するための学習部3301の処理の詳細については後述する。
(学習部3301)
学習部3301は、過去の顧客・契約データを学習データとして利用し、予測モデルの各種パラメータを機械学習によって算出(生成)する処理を行う。
上述したように、例えば年額定額制動画配信Webサービスの離脱防止マーケティング施策を例にして、年額定額制動画配信Webサービスが1年ごとの契約制になっている場合、学習データ3102は、例えば過去11か月に契約更新を行った又は離脱した顧客に対して生成される。つまり、予測対象となる今後1カ月以内に契約更新時期を迎える顧客は含まれない。
学習部3301は、学習データ3102を用いて機械学習を実施し、各種パラメータを算出(生成)する生成部として機能する。例えば学習部3301は、予測部3302による施策成功確率予測処理に用いられる確率予測モデルに使用する各種パラメータを機械学習により算出する。以下、学習部3301による確率予測モデルの学習について具体的に説明する。なお以下に説明する学習方法は一例であって、必ずしもこれに限定されない。
予測部3306は、学習部3301で学習したパラメータを利用し、予測モデルによって「施策を実施した場合の継続確率-施策を実施しない場合の継続確率」を予測する。
(xi;wy2)-(1-fh(x;wh))(つまり、「施策を実施した場合の継続確率-施策を実施しない場合の継続確率」)ではなく、fh(x;wh)×fy1(xi;wy1)+(1-fh(x;wh))×fy2(xi;wy2)-(1-fh(x;wh))-(fy1(xi;wy1)の予測確信度+fy2(xi;wy2)の予測確信度)の順に施策を実施する。
Recommendation,” Lihong Li, Wei Chu, John Langford, and Robert E. Schapire,
Proceedings of the 19th International Conference on World Wide Web, Pages 661‐670, 2010
算出部3303は、予測部3302によって出力された確率を用いて、クライアント100のユーザI/F部1401から要求された数値を算出する。また、算出部3303は、クライアント100の入出力部140に含まれるディスプレイなどのユーザI/F部1401に出力させるためのデータを生成し、通信部320からクライアント100に送信させる。かかるデータは、例えば算出した数値を含む結果画像である。なお、クライアント100における情報出力の方法は画像表示には限られず、例えば音声出力が画像表示と共に、またはこれに代えて採用されてもよい。
続いて、本実施形態による動作処理について、図6~図9を参照して具体的に説明する。
図6は、本実施形態による学習部3301の動作処理を示すフローチャートである。図6に示すように、まず、学習部3301は、顧客・契約データに基づいて、過去11か月に契約更新を行った又は離脱した顧客を選択し(ステップS103)、選択された顧客について顧客データエントリを作成し、その集合を学習データ3102とする(ステップS106)。顧客データエントリは、上述したように、3つの要素、「契約中か否か」、「施策を実施したか否か」、および「特徴量ベクトル」から成る。「特徴量ベクトル」の作成処理については、図7を参照して後述する。
図7は、学習部3301による特徴量ベクトルの作成処理を示すフローチャートである。図7に示すように、学習部3301は、顧客情報と契約情報の各項目について、one-of-kベクトルを作成する(ステップS123)。one-of-kベクトルの作成処理については、図8を参照して後述する。
図8は、本実施形態によるone-of-kベクトルの作成処理を示すフローチャートである。図8に示すように、学習部3301は、顧客情報、契約情報から、項目Aの値X、という情報を取得する(ステップS133)。
図9は、本実施形態による予測部3302の動作処理を示すフローチャートである。図9に示すように、まず、予測部3302は、今後1カ月以内に契約更新時期を迎える顧客を選択する(ステップS143)。
続いて、例えばクライアント100において入出力部140に含まれるディスプレイ(ユーザI/F部1401)に表示される画面の例を参照しながら、本開示の一実施形態において提示される情報の例について説明する。なお、以下の説明では年額定額制動画配信サービスの離脱防止マーケティング施策を実施する対象顧客を選択するにあたって提示される情報の例が説明されるが、他の契約の離脱防止マーケティング施策を実施する対象顧客を選択するにあたっても同様に情報が提示されうる。
・・・式3
・・・式4
・コスト:施策1件当たりの費用
・期待利益:{(顧客利益の予測値)-(クーポンの割引額)}×((施策を実施した場合の継続確率-施策を実施しない場合の継続確率)
・期待効果:{(顧客利益の予測値)-(クーポンの割引額)}×(施策を実施した場合の継続確率-施策を実施しない場合の継続確率)-(施策1件当たりのコスト)
(xi;wy2 )を用いて、「離脱防止確率」を施策の種類毎に算出する。かかるモデルの尤度の算出式を図14に示す。
次に、図16を参照して、本開示の実施形態による情報処理装置のハードウェア構成について説明する。図16は、本開示の実施形態による情報処理装置のハードウェア構成例を示すブロック図である。図示された情報処理装置900は、例えば、上記の実施形態におけるサーバ300およびクライアント100を実現しうる。
本開示の実施形態は、例えば、上記で説明したような情報処理装置(サーバまたはクライアント)、システム、情報処理装置またはシステムで実行される情報処理方法、情報処理装置を機能させるためのプログラム、およびプログラムが記録された一時的でない有形の媒体を含みうる。
・オンラインバンクなどのWebサイト経由で継続的に利用するサービスの継続促進、上位契約の促進、他サービスの契約促進
・オンライン学習サイトへの訪問促進、利用時間継続促進
・カーシェアリングなどシェアリングサービスの利用促進、比較的空いている時間の利用促進
・イーコマースサイトへの訪問促進、抱き合わせ販売の促進
・学習塾への通学促進
・飲食店の定期的な利用の継続促進
・病院への継続的な通院を促進
・老人ホームへの家族の訪問を促進
・物品等のレンタルサービスのレンタル期間延長を促進、他の製品のレンタルを促進
(1)学習データにおける類似度が一定以上の顧客数または類似上位一定人数の類似度の平均(この場合、数値が高いほど予測の確信度が高いと言える)
(2)学習データにおける類似顧客における離脱または継続のばらつきの大きさ(この場合、ばらつきが小さいほど予測の確信度が高いと言える)
(3)p2(xi)のパラメータの事後確率とxiの関係(この場合、xiの推定に使うパラメータのベイズ事後確率が高いほど予測の確信度が高いと言える)
(1)
顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導くための施策が成功する指標を出力する出力部を備える、情報処理装置。
(2)
前記出力部は、現在継続状態の顧客毎に、前記目的の状態へ導くための施策が成功する指標として施策成功確率を算出し、出力する、前記(1)に記載の情報処理装置。
(3)
前記顧客の継続状態は、契約状態であって、
前記継続状態の変化は、契約の離脱であって、
前記目的の状態は、前記契約状態を維持する状態である、前記(2)に記載の情報処理装置。
(4)
前記情報処理装置は、
前記状態に関する過去情報と、前記施策の実施に関する過去情報と、前記顧客または顧客の継続状態に関する特徴量とに基づいて、予測モデルのパラメータを生成する学習部と、
前記生成したパラメータを用いて、現在継続状態の顧客における、前記施策を実施しない場合の契約離脱の確率、および前記施策を実施した場合の契約離脱防止の確率を算出する予測部と、
を備え、
前記出力部は、前記施策を実施しない場合の契約離脱の確率と前記施策を実施した場合の契約離脱防止の確率の積を、前記施策が成功する指標として現在継続状態の顧客毎に算出する、前記(3)に記載の情報処理装置。
(5)
前記学習部は、前記施策を実施した場合の契約離脱防止の確率について前記施策を行わずとも契約離脱防止できた確率を考慮した予測モデルのパラメータを生成する、前記(4)に記載の情報処理装置。
(6)
前記出力部は、前記積を、前記施策が成功する指標を示すスコアとして、当該スコアの順に顧客IDを並べた一覧をユーザに出力する、前記(4)または(5)に記載の情報処理装置。
(7)
前記予測部は、予測の確信度を算出し、
前記出力部は、前記スコアを予測の確信度と併せて出力する、前記(6)に記載の情報処理装置。
(8)
前記出力部は、契約維持時の収益予想と前記施策の費用とに基づいて、施策実施時の利益率を算出し、ユーザに出力する、前記(4)~(7)のいずれか1項に記載の情報処理装置。
(9)
前記出力部は、前記利益率の予想グラフをユーザに出力する、前記(8)に記載の情報処理装置。
(10)
前記出力部は、施策を実施した場合の契約離脱防止の確率に対してプラスに働く特徴とマイナスに働く特徴をユーザに提示する、前記(4)~(9)のいずれか1項に記載の情報処理装置。
(11)
前記顧客の継続状態は、契約状態であって、
前記継続状態の変化は、契約のランクアップであって、
前記目的の状態は、前記契約状態がランクアップする状態である、前記(1)または(2)に記載の情報処理装置。
(12)
プロセッサが、
顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導く施策が成功する指標を出力するよう制御することを含む、情報処理方法。
(13)
コンピュータを、
顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導く施策が成功する指標を出力するよう制御する出力制御部として機能させるための、プログラム。
100 クライアント
110 ローカルストレージ
120 通信部
130 処理部
1301 算出部
140 入出力部
1401 ユーザI/F部
200 ネットワーク
300 サーバ
310 データベース
3101 顧客・契約データ
3301 学習部
3302 予測部
3303 算出部
320 通信部
330 処理部
Claims (13)
- 顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導くための施策が成功する指標を出力する出力部を備える、情報処理装置。
- 前記出力部は、現在継続状態の顧客毎に、前記目的の状態へ導くための施策が成功する指標として施策成功確率を算出し、出力する、請求項1に記載の情報処理装置。
- 前記顧客の継続状態は、契約状態であって、
前記継続状態の変化は、契約の離脱であって、
前記目的の状態は、前記契約状態を維持する状態である、請求項2に記載の情報処理装置。 - 前記情報処理装置は、
前記状態に関する過去情報と、前記施策の実施に関する過去情報と、前記顧客または顧客の継続状態に関する特徴量とに基づいて、予測モデルのパラメータを生成する学習部と、
前記生成したパラメータを用いて、現在継続状態の顧客における、前記施策を実施しない場合の契約離脱の確率、および前記施策を実施した場合の契約離脱防止の確率を算出する予測部と、
を備え、
前記出力部は、前記施策を実施しない場合の契約離脱の確率と前記施策を実施した場合の契約離脱防止の確率の積を、前記施策が成功する指標として現在継続状態の顧客毎に算出する、請求項3に記載の情報処理装置。 - 前記学習部は、前記施策を実施した場合の契約離脱防止の確率について前記施策を行わずとも契約離脱防止できた確率を考慮した予測モデルのパラメータを生成する、請求項4に記載の情報処理装置。
- 前記出力部は、前記積を、前記施策が成功する指標を示すスコアとして、当該スコアの順に顧客IDを並べた一覧をユーザに出力する、請求項4に記載の情報処理装置。
- 前記予測部は、予測の確信度を算出し、
前記出力部は、前記スコアを予測の確信度と併せて出力する、請求項6に記載の情報処理装置。 - 前記出力部は、契約維持時の収益予想と前記施策の費用とに基づいて、施策実施時の利益率を算出し、ユーザに出力する、請求項4に記載の情報処理装置。
- 前記出力部は、前記利益率の予想グラフをユーザに出力する、請求項8に記載の情報処理装置。
- 前記出力部は、施策を実施した場合の契約離脱防止の確率に対してプラスに働く特徴とマイナスに働く特徴をユーザに提示する、請求項4~9のいずれか1項に記載の情報処理装置。
- 前記顧客の継続状態は、契約状態であって、
前記継続状態の変化は、契約のランクアップであって、
前記目的の状態は、前記契約状態がランクアップする状態である、請求項1または2に記載の情報処理装置。 - プロセッサが、
顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導く施策が成功する指標を出力するよう制御することを含む、情報処理方法。 - コンピュータを、
顧客の継続状態の変化を示す過去情報と、変化が生じた顧客に対して目的の状態へ導く施策を実施したか否かを示す過去情報と、に基づいて算出された、前記目的の状態へ導く施策が成功する指標を出力するよう制御する出力制御部として機能させるための、プログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/069,089 US20190019202A1 (en) | 2016-01-29 | 2016-10-25 | Information processing apparatus, information processing method, and program |
EP16888073.0A EP3410379A1 (en) | 2016-01-29 | 2016-10-25 | Information processing device, information processing method, and program |
JP2017563688A JP6848883B2 (ja) | 2016-01-29 | 2016-10-25 | 情報処理装置、情報処理方法、およびプログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2016016608 | 2016-01-29 | ||
JP2016-016608 | 2016-01-29 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2017130488A1 true WO2017130488A1 (ja) | 2017-08-03 |
Family
ID=59398791
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2016/081598 WO2017130488A1 (ja) | 2016-01-29 | 2016-10-25 | 情報処理装置、情報処理方法、およびプログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US20190019202A1 (ja) |
EP (1) | EP3410379A1 (ja) |
JP (1) | JP6848883B2 (ja) |
WO (1) | WO2017130488A1 (ja) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2019066698A (ja) * | 2017-10-02 | 2019-04-25 | 株式会社増進会ホールディングス | 学習継続システムおよび学習継続方法 |
CN111260466A (zh) * | 2020-01-20 | 2020-06-09 | 北京合信力科技有限公司 | 一种触达任务的处理方法及装置 |
JP2020095469A (ja) * | 2018-12-12 | 2020-06-18 | 富士通株式会社 | 離反防止支援プログラム、離反防止支援方法および情報処理装置 |
JP2020177262A (ja) * | 2019-04-15 | 2020-10-29 | ヤフー株式会社 | 情報処理装置、システム、情報処理方法、およびプログラム |
JP6816905B1 (ja) * | 2020-01-23 | 2021-01-20 | 株式会社クロス・デジタル・イノベーション | ユーザの個人情報の売買を支援するためのシステム、方法、およびプログラム |
JP2021111007A (ja) * | 2020-01-07 | 2021-08-02 | 株式会社アジェンダ | 脱退見込み評価装置 |
WO2024047812A1 (ja) * | 2022-08-31 | 2024-03-07 | 三菱電機株式会社 | 情報処理装置、出力方法、及び出力プログラム |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11276046B2 (en) * | 2018-10-16 | 2022-03-15 | Dell Products L.P. | System for insights on factors influencing payment |
JP2020087023A (ja) * | 2018-11-27 | 2020-06-04 | 日本電信電話株式会社 | 受注予測モデルの生成方法、受注予測モデル、受注予測装置、受注予測方法および受注予測プログラム |
US11861732B1 (en) * | 2022-07-27 | 2024-01-02 | Intuit Inc. | Industry-profile service for fraud detection |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002269329A (ja) * | 2001-01-05 | 2002-09-20 | Yasufumi Uchiumi | 業務改善支援システムおよびその方法 |
JP2015014828A (ja) * | 2013-07-03 | 2015-01-22 | 日本電信電話株式会社 | 離脱推定装置、解約推定装置、方法、及びプログラム |
JP2015166989A (ja) * | 2014-03-04 | 2015-09-24 | 株式会社野村総合研究所 | 情報処理装置および情報分析方法 |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7340408B1 (en) * | 2000-06-13 | 2008-03-04 | Verizon Laboratories Inc. | Method for evaluating customer valve to guide loyalty and retention programs |
KR20030086249A (ko) * | 2001-01-05 | 2003-11-07 | 야스후미 우츠미 | 업무개선지원 시스템 및 그 방법 |
US8712828B2 (en) * | 2005-12-30 | 2014-04-29 | Accenture Global Services Limited | Churn prediction and management system |
WO2011112173A1 (en) * | 2010-03-08 | 2011-09-15 | Hewlett-Packard Development Company, L.P. | Methods and systems for identifying customer status for developing customer retention and loyalty strategies |
US8630892B2 (en) * | 2011-08-31 | 2014-01-14 | Accenture Global Services Limited | Churn analysis system |
US9232063B2 (en) * | 2013-10-31 | 2016-01-05 | Verint Systems Inc. | Call flow and discourse analysis |
US11488086B2 (en) * | 2014-10-13 | 2022-11-01 | ServiceSource International, Inc. | User interface and underlying data analytics for customer success management |
-
2016
- 2016-10-25 EP EP16888073.0A patent/EP3410379A1/en not_active Ceased
- 2016-10-25 JP JP2017563688A patent/JP6848883B2/ja active Active
- 2016-10-25 WO PCT/JP2016/081598 patent/WO2017130488A1/ja active Application Filing
- 2016-10-25 US US16/069,089 patent/US20190019202A1/en not_active Abandoned
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002269329A (ja) * | 2001-01-05 | 2002-09-20 | Yasufumi Uchiumi | 業務改善支援システムおよびその方法 |
JP2015014828A (ja) * | 2013-07-03 | 2015-01-22 | 日本電信電話株式会社 | 離脱推定装置、解約推定装置、方法、及びプログラム |
JP2015166989A (ja) * | 2014-03-04 | 2015-09-24 | 株式会社野村総合研究所 | 情報処理装置および情報分析方法 |
Non-Patent Citations (1)
Title |
---|
See also references of EP3410379A4 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2019066698A (ja) * | 2017-10-02 | 2019-04-25 | 株式会社増進会ホールディングス | 学習継続システムおよび学習継続方法 |
JP2020095469A (ja) * | 2018-12-12 | 2020-06-18 | 富士通株式会社 | 離反防止支援プログラム、離反防止支援方法および情報処理装置 |
JP7215130B2 (ja) | 2018-12-12 | 2023-01-31 | 富士通株式会社 | 離反防止支援プログラム、離反防止支援方法および情報処理装置 |
JP2020177262A (ja) * | 2019-04-15 | 2020-10-29 | ヤフー株式会社 | 情報処理装置、システム、情報処理方法、およびプログラム |
JP7043453B2 (ja) | 2019-04-15 | 2022-03-29 | ヤフー株式会社 | 情報処理装置、システム、情報処理方法、およびプログラム |
JP2021111007A (ja) * | 2020-01-07 | 2021-08-02 | 株式会社アジェンダ | 脱退見込み評価装置 |
JP7037708B2 (ja) | 2020-01-07 | 2022-03-17 | 株式会社アジェンダ | 脱退見込み評価装置 |
CN111260466A (zh) * | 2020-01-20 | 2020-06-09 | 北京合信力科技有限公司 | 一种触达任务的处理方法及装置 |
JP6816905B1 (ja) * | 2020-01-23 | 2021-01-20 | 株式会社クロス・デジタル・イノベーション | ユーザの個人情報の売買を支援するためのシステム、方法、およびプログラム |
WO2021149804A1 (ja) * | 2020-01-23 | 2021-07-29 | 株式会社クロス・デジタル・イノベーション | ユーザの個人情報の売買を支援するためのシステム、方法、およびプログラム |
JP2021117598A (ja) * | 2020-01-23 | 2021-08-10 | 株式会社クロス・デジタル・イノベーション | ユーザの個人情報の売買を支援するためのシステム、方法、およびプログラム |
WO2024047812A1 (ja) * | 2022-08-31 | 2024-03-07 | 三菱電機株式会社 | 情報処理装置、出力方法、及び出力プログラム |
Also Published As
Publication number | Publication date |
---|---|
US20190019202A1 (en) | 2019-01-17 |
EP3410379A4 (en) | 2018-12-05 |
JP6848883B2 (ja) | 2021-03-24 |
EP3410379A1 (en) | 2018-12-05 |
JPWO2017130488A1 (ja) | 2018-11-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2017130488A1 (ja) | 情報処理装置、情報処理方法、およびプログラム | |
JP6719727B2 (ja) | 購買行動分析装置およびプログラム | |
JP6767824B2 (ja) | 判定装置、判定方法及び判定プログラム | |
US11188950B2 (en) | Audience expansion for online social network content | |
US20150006286A1 (en) | Targeting users based on categorical content interactions | |
US20150006294A1 (en) | Targeting rules based on previous recommendations | |
US10783431B2 (en) | Image search using emotions | |
JP2010250827A (ja) | タッチポイントをカスタマイズするシステム | |
US20210390610A1 (en) | Multi-layer optimization for a multi-sided network service | |
CN102365649A (zh) | 为了广告的推断定向而利用社交网络中的信息 | |
JP2018045600A (ja) | 判定装置、判定方法、及び判定プログラム | |
EP3437055A2 (en) | Method and system for determining optimized customer touchpoints | |
JP2018045570A (ja) | 生成装置、生成方法、及び生成プログラム | |
JP7348230B2 (ja) | 生成装置、生成方法、及び生成プログラム | |
JP2020057221A (ja) | 情報処理方法、情報処理装置及びプログラム | |
JP6241698B1 (ja) | 評価支援システム及び評価支援装置 | |
JP6288748B1 (ja) | 評価支援システム及び評価支援装置 | |
JP6494576B2 (ja) | 推定装置、推定方法、及び推定プログラム | |
US20160148271A1 (en) | Personalized Marketing Based on Sequence Mining | |
WO2020084712A1 (ja) | 配信プログラム、配信方法及び配信装置 | |
JP2020154464A (ja) | 推定装置、推定方法及び推定プログラム | |
Pamuru et al. | Status Downgrade: The Impact of Losing Status on a User-generated Content Platform | |
JP7434264B2 (ja) | 情報処理装置、情報処理方法、及び情報処理プログラム | |
JP7416904B1 (ja) | 情報処理装置、情報処理方法、および情報処理プログラム | |
JP7459040B2 (ja) | 情報処理装置、情報処理方法、及び情報処理プログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16888073 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2017563688 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2016888073 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2016888073 Country of ref document: EP Effective date: 20180829 |