CN116049666A - Customer loss prediction method, device, computer equipment and storage medium - Google Patents
Customer loss prediction method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN116049666A CN116049666A CN202211615580.6A CN202211615580A CN116049666A CN 116049666 A CN116049666 A CN 116049666A CN 202211615580 A CN202211615580 A CN 202211615580A CN 116049666 A CN116049666 A CN 116049666A
- Authority
- CN
- China
- Prior art keywords
- data
- clients
- customer
- predicted
- loss
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present disclosure relates to the field of customer loss prediction analysis technologies, and in particular, to a customer loss prediction method, a device, a computer device, and a storage medium. A customer churn prediction method includes obtaining a data set of a customer to be predicted, the data set including at least one of consumption data and recharge data; according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients; and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted. The method establishes a decision tree model based on the historical data of the clients, evaluates and timely early warns the loss risk of the clients, is not limited by industry and operation modes, can accurately position the loss clients, and further carries out targeted recall on the clients, so that the marketing cost is reduced.
Description
Technical Field
The present disclosure relates to the field of customer loss prediction analysis technologies, and in particular, to a customer loss prediction method, a device, a computer device, and a storage medium.
Background
In the related technical field, the customer loss prediction method is to construct a basic portrait for the customers according to basic label data of the customers, and then to predict the value grouping loss risk of the customers by means of an analysis model, so as to improve the profit and the income of targeted marketing of different customers. The client portrait is a labeled client model abstracted according to the information of the client such as the attribute, the client preference, the living habit, the client behavior and the like, the core work of constructing the client portrait is to label the client, and the label is the interpretation and the high extraction of the client behavior, the attribute and the like. The portrait tag systems of different services are respectively heavy and can be continuously enriched according to data and operation purposes. Along with the continuous accumulation of huge data, continuous enrichment of business scenes and continuous optimization of algorithm models, customer portraits can become comprehensive and accurate.
The client image is used for predicting the client loss, so that the dependence on business experience is high, and particularly the client loss prediction under a specific marketing mode is high in limitation and low in accuracy.
Therefore, how to accurately predict the loss of the client is a technical problem to be solved by the technicians in the field without being influenced by the restrictions of industry and operation modes.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a customer churn prediction method, apparatus, computer device, and storage medium.
In a first aspect, the present application provides a customer churn prediction method including:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In one embodiment, determining the clients to be predicted as the lost clients and the non-lost clients based on the consumption data and the recharge data includes:
determining clients to be predicted, which have no consumption data and no recharging data within a preset period, as lost clients;
and determining the clients to be predicted, which have consumption data and recharging data in a preset period, as non-loss clients.
In one embodiment, the method further includes, before performing loss prediction on the non-loss clients through the pre-trained decision tree model to obtain loss pre-warning information of the clients to be predicted:
performing discrete value conversion on numerical data in the data set;
and performing independent discrete conversion on attribute value data which are mutually independent in the data set.
In one embodiment, the method further includes, before performing the attrition prediction on the non-attrition client through the pre-established decision tree model to obtain the attrition behavior prediction information of the client to be predicted:
and calculating the information gain rate of each data of the data set, and obtaining basic attribute data according to the information gain rate and the threshold value.
In one embodiment, building a decision tree model includes:
taking basic attribute data with the maximum information gain rate as a root node;
establishing a tree bifurcation for each value of the root node;
selecting a sample subset from the basic attribute data for each bifurcation, and establishing nodes for the rest basic attribute data;
the above procedure is iterated until there is no remaining basic attribute data, which node is defined as a leaf node.
In one embodiment, establishing the decision tree model further comprises:
deleting the subtrees taking the leaf nodes as the roots by the leaf nodes without the samples;
the classification of the data set associated with the leaf node is given.
In one embodiment, the method performs churn prediction on non-churn clients through a pre-established decision tree model to obtain churn behavior prediction information of clients to be predicted, and further includes:
and calculating the loss function value of the decision tree model through the test data set.
In a second aspect, the present application further provides a customer churn prediction apparatus, including: the customer churn prediction apparatus includes:
the system comprises an acquisition unit, a prediction unit and a storage unit, wherein the acquisition unit is used for acquiring a data set of a client to be predicted, and the data set comprises at least one of consumption data and recharging data;
the customer analysis unit to be predicted is used for determining the customer to be predicted as an already lost customer and a non-lost customer according to the consumption data and the recharging data;
and the loss behavior prediction unit is used for carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In a fourth aspect, the present application also provides a computer-readable storage medium. A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
The method, the device, the computer equipment and the storage medium for predicting the customer loss are characterized in that a data set of a customer to be predicted is obtained, wherein the data set comprises at least one of consumption data and recharging data; according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients; and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted. The method establishes a decision tree model based on the historical data of the clients, evaluates and timely early warns the loss risk of the clients, is not limited by industry and operation modes, can accurately position the loss clients, and further carries out targeted recall on the clients, so that the marketing cost is reduced.
Drawings
In order to more clearly illustrate the technical solutions of embodiments or conventional techniques of the present application, the drawings required for the descriptions of the embodiments or conventional techniques will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person of ordinary skill in the art.
FIG. 1 is a flow chart of a client churn prediction method according to one embodiment;
FIG. 2 is a flow diagram of a method for customer churn prediction in one embodiment for determining churn customers and non-churn customers;
FIG. 3 is a schematic diagram of a data processing flow of a client churn prediction method according to another embodiment;
FIG. 4 is a flow chart of a decision tree model building method for client churn prediction in one embodiment;
fig. 5 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
In order to facilitate an understanding of the present application, a more complete description of the present application will now be provided with reference to the relevant figures. Examples of the present application are given in the accompanying drawings. This application may, however, be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and/or" as used herein includes any and all couplings of one or more of the associated listed items.
It will be understood that the terms "first," "second," and the like, as used herein, may be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another element.
The client loss prediction method provided by the embodiment of the application can be applied to the marketing field, and the client loss prediction method can predict the client loss tendency in advance by establishing a regression model based on a decision tree, is convenient for early intervention, and comprises the steps of acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data; according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients; and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted. The method establishes a decision tree model based on the historical data of the clients, evaluates and timely early warns the loss risk of the clients, is not limited by industry and operation modes, can accurately position the loss clients, and further carries out targeted recall on the clients, so that the marketing cost is reduced.
Embodiment 1,
As shown in fig. 1, in this embodiment, a client churn prediction method is provided, which includes the following steps:
s101: a dataset of customers to be predicted is obtained, the dataset comprising at least one of consumption data and refill data.
The clients to be predicted refer to clients to be subjected to attrition prediction.
Specifically, the customer churn prediction means acquires a data set of customers to be predicted, the data set including at least one of consumption data and refill data. The data set may be data within a predetermined period, such as half a year, or may be data within other periods. The dataset may include a customer information table, specific fields relate to: customer id, registration date, sex, age; the half consumption information record table of the client in the last year comprises the following specific fields: customer id, consumption time, consumption amount, whether to offer, offer amount, and commodity class; the half-recharging information record table of the client in the last year comprises the following specific fields: customer id, recharge time, recharge amount.
S102: and determining the clients to be predicted as lost clients and non-lost clients according to the consumption data and the recharging data.
Specifically, the customer loss prediction device divides the consumption data and the recharging data of the customer in the last year and half into the consumption data and the recharging data of the previous year and the consumption data and the recharging data of the second year according to the consumption data and the recharging data acquired in the step S101, if the consumption data and the recharging data information are not available in the last year of a certain customer, the customer is marked as a lost customer, otherwise, the customer is marked as a non-lost customer.
S103: and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
Specifically, the customer loss prediction device inputs consumption data and recharging data of a non-loss customer in half a year into a pre-established decision tree model, and carries out loss prediction on the non-loss customer through the pre-established decision tree model to obtain loss behavior prediction information of the customer to be predicted.
In this embodiment, a customer churn prediction method is provided by acquiring a data set of a customer to be predicted, the data set including at least one of consumption data and refill data; according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients; and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted. The method establishes a decision tree model based on the historical data of the clients, evaluates and timely early warns the loss risk of the clients, is not limited by industry and operation modes, can accurately position the loss clients, and further carries out targeted recall on the clients, so that the marketing cost is reduced.
Embodiment II,
As shown in fig. 2, in the present embodiment, step S102 is provided: according to the consumption data and the recharging data, the clients to be predicted are determined to be lost clients and non-lost clients, and the method comprises the following steps:
s1021: and determining the clients to be predicted, which have no consumption data and no recharging data within a preset period, as lost clients.
Specifically, the customer loss prediction apparatus divides the consumption data and the recharge data of the customer for the last half year into the consumption data and the recharge data of the previous year and the consumption data and the recharge data of the second half year according to the consumption data and the recharge data acquired in step S101. The method comprises the steps of processing and analyzing consumption data and recharging data of the previous year as basic data, such as date, client ID, gender, age, registration time, first transaction time, joining duration, last consumption time, last consumption times, last year consumption class number, last year average consumption amount, last year maximum consumption interval and last year average consumption interval; the recharging times of the last 1 year, the recharging amount of the last 1 year and the recharging amount of the last 1 year.
S1022: and determining the clients to be predicted, which have consumption data and recharging data in a preset period, as non-loss clients.
Specifically, the customer loss prediction device divides the consumption data and the recharge data of the customer in the last half year into the consumption data and the recharge data of the previous year and the consumption data and the recharge data of the second half year. The consumption data and the recharging data of the last half year are used as data for marking whether customers run off, such as the consumption times of the last half year, the consumption class number of the last half year, the average consumption amount of the last half year, the consumption count of the last 30 days, the consumption amount of the last 30 days, the preferential amount of the last 30 days, the recharging times of the last half year, the recharging amount of the last half year, the recharging times of the last 30 days and the recharging amount of the last 30 days. And the consumption data and the recharging data of the last half year mark a client as lost if the recharging and consumption information of the client is not available in the last half year, and mark no loss otherwise.
In this embodiment, a customer churn prediction method is provided, in which a customer churn prediction device determines a customer to be predicted, which has no consumption data and no recharge data for a preset period, as a churn customer; and determining the clients to be predicted, which have consumption data and recharging data in a preset period, as non-loss clients. The customer loss prediction device marks the lost and non-lost customers according to the service definition, the consumption data and the recharging data of the customers in the last half year, and can also directly mark the lost according to other service definitions. The lost clients can be marked, so that the dependence of the decision tree model on the historical data quantity can be reduced, the data processing quantity is reduced, and the working efficiency and the prediction accuracy are improved.
Third embodiment,
As shown in fig. 3, in the present embodiment, step S103 is provided: the method comprises the following steps of:
s1031: discrete value conversion is performed on the numerical data in the data set.
Specifically, the customer churn prediction device also normalizes the data before churn prediction is performed on non-churn customers. The customer loss prediction device performs discrete value conversion on the numerical attribute values such as the amount of money, the times, the days and the like in the data set according to 30-bit, 50-bit and 70-bit values respectively.
S1032: and performing independent discrete conversion on attribute value data which are mutually independent in the data set.
Specifically, the customer churn prediction device also normalizes the data before churn prediction is performed on non-churn customers. The customer churn prediction device performs irrelevant discrete conversion on the attribute value data such as name and sex which are mutually independent in the set by using one-hot coding to obtain opedate, userid, sex and age.
In this embodiment, the customer churn prediction means also normalizes the data before churn prediction is performed on non-churn customers. The customer loss prediction device performs discrete value conversion on the numerical data in the data set, performs irrelevant discrete conversion on the mutually independent attribute value data in the data set, and obtains standardized data, so that the prediction of the decision tree model is facilitated, and the accuracy is further improved.
Fourth embodiment,
In the present embodiment, step S103 is provided: carrying out loss prediction on non-loss clients through a pre-established decision tree model, and before obtaining loss behavior prediction information of the clients to be predicted, comprising the following steps:
and calculating the information gain rate of each data of the data set, and obtaining basic attribute data according to the information gain rate and the threshold value.
Specifically, the customer churn prediction device calculates the information gain rate of each data in the data set, evaluates the importance of these attributes by using the information gain parameters, and classifies the expected information of the attributes arbitrarily:
ordering { a } of m attribute values of a certain continuous attribute A according to natural sequence 1 ，a 2 ，…，a m Taking the mean value of two adjacent points as the division points to obtain m-1 division points of the group of attributes, wherein the ith division point is:
for each division point T i Wherein i.epsilon.1, m-1]Calculated as T i Information gain maximum value as classification point:
wherein j is E [1, m-2 ]]The discrete value of the continuous attribute A is a x ＝1，a y =2, where x∈ [1, j ]]，y∈[j+1，m]I.e. a discretization of the continuous properties is achieved.
Then, the importance of these attributes is evaluated by using the information gain parameters, and the expected information of the attributes is arbitrarily classified:
wherein A represents a certain attribute, n represents n values of A, p i The probability of the i-th value. H (a) represents desired information.
The entropy divided into subsets by A is calculated as follows:
where A represents an attribute, representing a division of the data set D by attribute A. Customer churn predictionThe device obtains basic attribute data according to the calculated information gain rate and the threshold value. H A (D) Representing the gain of information g R (D, A) represents the information gain ratio.
As shown in fig. 4, in the present embodiment, there is provided a decision tree model establishment including:
s201: and taking the basic attribute data with the maximum information gain rate as a root node.
S202: a tree bifurcation is established for each value of the root node.
S203: and selecting a sample subset from the basic attribute data for each bifurcation, and establishing nodes for the rest basic attribute data.
S204: the above procedure is iterated until there is no remaining basic attribute data, which node is defined as a leaf node.
S205: and deleting the subtrees with the leaf nodes as the root by the leaf nodes without the samples.
S206: the classification of the data set associated with the leaf node is given.
In this embodiment, the customer loss prediction apparatus establishes a decision tree model, so that the decision tree model can process continuity features, and the deviation problem of the algorithm on the multi-value attribute can be corrected by using the information gain rate as an evaluation standard, so that feature missing data can be processed, and the over-fitting problem is avoided.
Fifth embodiment (V),
In the present embodiment, step S103 is provided: the method comprises the following steps of: and calculating the loss function value of the decision tree model through the test data set.
Specifically, the customer churn prediction apparatus calculates a churn function value of the decision tree model from the test dataset. The customer churn prediction means follows the data set as 7:3, randomly dividing the model into a training set and a testing set, wherein the training set is used for training the decision tree model, and the testing set is used for testing the prediction accuracy of the trained decision tree model. The accuracy of the algorithm is defined by the accuracy of the prediction of the is-loss value of the test set through the algorithm, and finally whether the customer runs off or not is predicted through the established decision tree.
The algorithm predicts the future loss behavior by using the historical data, and the loss label in the model is marked according to the service definition and the current behavior of the client, so that more historical data is needed to ensure the accuracy of the model; the method can also define the loss marks directly according to the business, so that the dependence of the model on the historical data amount can be reduced, but the accuracy of the model depends more on business experience, and for a fixed business scene, the latter is a more efficient model of a row, and the former is a more general algorithm model on the premise of ensuring the accuracy rate
It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a customer loss prediction device for implementing the above-mentioned customer loss prediction method. The implementation of the solution provided by the customer loss prediction apparatus is similar to that described in the above method, so the following specific limitation of one or more embodiments of the customer loss prediction apparatus may be referred to the above limitation of the customer loss prediction method, and will not be repeated herein.
In one embodiment, a customer churn prediction apparatus is provided, comprising:
the system comprises an acquisition unit, a prediction unit and a storage unit, wherein the acquisition unit is used for acquiring a data set of a client to be predicted, and the data set comprises at least one of consumption data and recharging data;
the customer analysis unit to be predicted is used for determining the customer to be predicted as an already lost customer and a non-lost customer according to the consumption data and the recharging data;
and the loss behavior prediction unit is used for carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In one embodiment, determining the customers to be predicted as the lost customers and the non-lost customers based on the consumption data and the refill data comprises:
the lost client analysis unit is used for determining clients to be predicted, which have no consumption data and no recharging data in a preset period, as lost clients;
and the non-attrition client unit is used for determining clients to be predicted, which have consumption data and recharging data in a preset period, as non-attrition clients.
In one embodiment, the method further includes, before performing loss prediction on the non-loss clients through the pre-trained decision tree model to obtain loss pre-warning information of the clients to be predicted:
the discrete value conversion unit is used for carrying out discrete value conversion on the numerical data in the data set;
and the independence discrete conversion unit is used for performing independence discrete conversion on the attribute value data which are mutually independent in the data set.
In one embodiment, the method further includes, before performing the attrition prediction on the non-attrition client through the pre-established decision tree model to obtain the attrition behavior prediction information of the client to be predicted:
and the basic attribute data unit is used for calculating the information gain rate of each data of the data set and obtaining basic attribute data according to the information gain rate and the threshold value.
In one embodiment, building a decision tree model includes:
the root node establishing unit is used for taking the basic attribute data with the maximum information gain rate as a root node;
a bifurcation establishing unit, configured to establish bifurcation of a tree for each value of the root node;
the node establishing unit is used for selecting a sample subset from the basic attribute data for each bifurcation and establishing nodes for the rest basic attribute data;
a leaf node unit for recursively repeating the above procedure until there is no remaining basic attribute data, the node being defined as a leaf node.
In one embodiment, establishing the decision tree model further comprises:
a deleting unit, configured to delete a subtree with a leaf node as a root, where the leaf node has no sample;
and the classification unit is used for giving classification to the data set associated with the leaf node.
In one embodiment, the method performs churn prediction on the non-churn clients through a pre-established decision tree model to obtain churn behavior prediction information of the clients to be predicted, and further includes:
and the loss function unit is used for calculating a loss function value of the decision tree model through the test data set.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 5. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used to store periodic task allocation data such as configuration files, theoretical operating parameters and theoretical deviation value ranges, task attribute information, and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program when executed by a processor implements a customer churn prediction method.
It will be appreciated by those skilled in the art that the structure shown in fig. 5 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory storing a computer program and a processor that when executing the computer program performs the steps of:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In one embodiment, the processor, when executing the computer program, implements determining clients to be predicted as lost clients and non-lost clients based on the consumption data and the refill data, comprising:
determining clients to be predicted, which have no consumption data and no recharging data within a preset period, as lost clients;
and determining the clients to be predicted, which have consumption data and recharging data in a preset period, as non-loss clients.
In one embodiment, when the processor executes the computer program, the method further includes, before implementing the churn prediction for the non-churn client through the pre-trained decision tree model to obtain churn pre-warning information of the client to be predicted:
performing discrete value conversion on numerical data in the data set;
and performing independent discrete conversion on attribute value data which are mutually independent in the data set.
In one embodiment, before implementing the attrition prediction for the non-attrition client through the pre-established decision tree model when the processor executes the computer program to obtain the attrition behavior prediction information of the client to be predicted, the method further includes:
and calculating the information gain rate of each data of the data set, and obtaining basic attribute data according to the information gain rate and the threshold value.
In one embodiment, a processor, when executing a computer program, implements building a decision tree model comprising:
taking basic attribute data with the maximum information gain rate as a root node;
establishing a tree bifurcation for each value of the root node;
selecting a sample subset from the basic attribute data for each bifurcation, and establishing nodes for the rest basic attribute data;
the above procedure is iterated until there is no remaining basic attribute data, which node is defined as a leaf node.
In one embodiment, the processor, when executing the computer program, implements building a decision tree model, further comprising:
deleting the subtrees taking the leaf nodes as the roots by the leaf nodes without the samples;
the classification of the data set associated with the leaf node is given.
In one embodiment, when the processor executes the computer program, the method performs attrition prediction on the non-attrition client through a pre-established decision tree model to obtain attrition behavior prediction information of the client to be predicted, and further includes:
and calculating the loss function value of the decision tree model through the test data set.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
In one embodiment, a computer program, when executed by a processor, enables determining clients to be predicted as lost clients and non-lost clients based on consumption data and refill data, comprising:
determining clients to be predicted, which have no consumption data and no recharging data within a preset period, as lost clients;
and determining the clients to be predicted, which have consumption data and recharging data in a preset period, as non-loss clients.
In one embodiment, the computer program when executed by the processor performs churn prediction on non-churn clients through a pre-trained decision tree model, and before obtaining churn pre-warning information of the clients to be predicted, the method further includes:
performing discrete value conversion on numerical data in the data set;
and performing independent discrete conversion on attribute value data which are mutually independent in the data set.
In one embodiment, the computer program when executed by the processor performs attrition prediction on the non-attrition client through a pre-established decision tree model, and before obtaining attrition behavior prediction information of the client to be predicted, the method further includes:
and calculating the information gain rate of each data of the data set, and obtaining basic attribute data according to the information gain rate and the threshold value.
In one embodiment, a computer program, when executed by a processor, implements building a decision tree model, comprising:
taking basic attribute data with the maximum information gain rate as a root node;
establishing a tree bifurcation for each value of the root node;
selecting a sample subset from the basic attribute data for each bifurcation, and establishing nodes for the rest basic attribute data;
the above procedure is iterated until there is no remaining basic attribute data, which node is defined as a leaf node.
In one embodiment, the computer program, when executed by the processor, implements the building of a decision tree model, further comprising:
deleting the subtrees taking the leaf nodes as the roots by the leaf nodes without the samples;
the classification of the data set associated with the leaf node is given.
In one embodiment, the computer program when executed by the processor performs attrition prediction on the non-attrition client through a pre-established decision tree model to obtain attrition behavior prediction information of the client to be predicted, and further includes:
and calculating the loss function value of the decision tree model through the test data set.
Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the various embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The various embodiments in this disclosure are described in a progressive manner, and identical and similar parts of the various embodiments are all referred to each other, and each embodiment is mainly described as different from other embodiments.
The scope of the present disclosure is not limited to the above-described embodiments, and it is apparent that various modifications and variations can be made to the present disclosure by those skilled in the art without departing from the scope and spirit of the disclosure. Such modifications and variations are intended to be included herein within the scope of the following claims and their equivalents.
Claims (10)
1. A customer churn prediction method, characterized in that the customer churn prediction method comprises:
acquiring a data set of a client to be predicted, wherein the data set comprises at least one of consumption data and recharging data;
according to the consumption data and the recharging data, determining the clients to be predicted as lost clients and non-lost clients;
and carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
2. The customer churn prediction method according to claim 1, wherein said determining the customer to be predicted as a churn customer and a non-churn customer based on the consumption data and the refill data comprises:
determining the clients to be predicted, which do not have the consumption data and the recharging data within a preset period, as lost clients;
and determining the clients to be predicted, which have the consumption data and the recharging data within a preset period, as non-loss clients.
3. The method for predicting customer churn according to claim 1, wherein said performing churn prediction on said non-churn customers through a pre-trained decision tree model, before obtaining churn pre-warning information of said customers to be predicted, further comprises:
performing discrete value conversion on numerical data in the data set;
and performing independent discrete conversion on attribute value data which are mutually independent in the data set.
4. The method for predicting customer churn according to claim 1, wherein before said churn prediction is performed on said non-churn customers through a pre-established decision tree model to obtain churn behavior prediction information of said customers to be predicted, further comprising:
and calculating the information gain rate of each data of the data set, and obtaining basic attribute data according to the information gain rate and a threshold value.
5. The customer churn prediction method of claim 4 wherein building the decision tree model comprises:
taking the basic attribute data with the maximum information gain rate as a root node;
establishing a tree bifurcation for each value of the root node;
selecting a sample subset from the basic attribute data for each bifurcation, and establishing nodes for the rest basic attribute data;
the above procedure is iterated until there are no remaining said basic attribute data, which node is defined as a leaf node.
6. The customer churn prediction method of claim 5 wherein building the decision tree model further comprises:
deleting the subtrees with the leaf nodes as roots from the leaf nodes without samples;
giving classification of the dataset with which the leaf node is associated.
7. The method for predicting customer churn according to any one of claims 1 to 6, wherein said churn prediction is performed on said non-churn customers through a pre-established decision tree model to obtain churn behavior prediction information of said customers to be predicted, further comprising:
and calculating a loss function value of the decision tree model through a test data set.
8. A customer churn prediction apparatus, characterized in that the customer churn prediction apparatus comprises:
an acquisition unit for acquiring a data set of a client to be predicted, the data set including at least one of consumption data and recharge data;
the customer analysis unit to be predicted is used for determining the customer to be predicted as an already lost customer and a non-lost customer according to the consumption data and the recharging data;
and the loss behavior prediction unit is used for carrying out loss prediction on the non-loss clients through a pre-established decision tree model to obtain loss behavior prediction information of the clients to be predicted.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.
10. A computer storage medium having stored thereon a computer program, which when executed by a processor realizes the steps of the method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211615580.6A CN116049666A (en) | 2022-12-15 | 2022-12-15 | Customer loss prediction method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211615580.6A CN116049666A (en) | 2022-12-15 | 2022-12-15 | Customer loss prediction method, device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116049666A true CN116049666A (en) | 2023-05-02 |
Family
ID=86132159
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211615580.6A Pending CN116049666A (en) | 2022-12-15 | 2022-12-15 | Customer loss prediction method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116049666A (en) |
-
2022
- 2022-12-15 CN CN202211615580.6A patent/CN116049666A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108229986B (en) | Feature construction method in information click prediction, information delivery method and device | |
| CN109886554B (en) | Illegal behavior discrimination method, device, computer equipment and storage medium | |
| CN108665293B (en) | Feature importance obtaining method and device | |
| CN109472310B (en) | Identification method and device for determining two resumes to be identical talents | |
| CN116362607B (en) | Material reserve efficiency evaluation method and device | |
| CN113326177A (en) | Index anomaly detection method, device, equipment and storage medium | |
| CN115204536A (en) | Building equipment fault prediction method, device, equipment and storage medium | |
| CN114491084B (en) | Self-encoder-based relation network information mining method, device and equipment | |
| CN113420887A (en) | Prediction model construction method and device, computer equipment and readable storage medium | |
| CN109325781B (en) | Customer quality analysis method, apparatus, computer device and storage medium | |
| CN112767032B (en) | Information processing method, device, electronic equipment and storage medium | |
| CN114254762A (en) | Interpretable machine learning model construction method and device and computer equipment | |
| CN114169942A (en) | User characterization method and system | |
| CN111950623B (en) | Data stability monitoring method, device, computer equipment and medium | |
| CN112348093A (en) | Data processing method and device, server and storage medium | |
| CN116049666A (en) | Customer loss prediction method, device, computer equipment and storage medium | |
| CN110874612B (en) | Time interval prediction method and device, computer equipment and storage medium | |
| CN112528662A (en) | Entity category identification method, device, equipment and storage medium based on meta-learning | |
| CN113592218A (en) | Photovoltaic user baseline load estimation method and device and terminal equipment | |
| CN114048825A (en) | Clustering method, system, equipment and storage medium based on contrast learning | |
| CN112927092A (en) | Method, device and equipment for predicting claim settlement trend and storage medium | |
| CN112328746A (en) | Dish label warehousing method and device, computer equipment and storage medium | |
| CN116611506B (en) | User analysis model training method, user label determining method and device | |
| CN112927000A (en) | Data analysis method and device for network points, computer equipment and storage medium | |
| CN109710876B (en) | Information recommendation method and device and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |