CN117788043A - Cloud computing industry cloud computer customer loss early warning method and system - Google Patents
Cloud computing industry cloud computer customer loss early warning method and system Download PDFInfo
- Publication number
- CN117788043A CN117788043A CN202311703144.9A CN202311703144A CN117788043A CN 117788043 A CN117788043 A CN 117788043A CN 202311703144 A CN202311703144 A CN 202311703144A CN 117788043 A CN117788043 A CN 117788043A
- Authority
- CN
- China
- Prior art keywords
- customer
- client
- loss
- information
- early warning
- 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
- 238000000034 method Methods 0.000 title claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 43
- 238000007781 pre-processing Methods 0.000 claims abstract description 18
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 238000004458 analytical method Methods 0.000 claims abstract description 7
- 238000007405 data analysis Methods 0.000 claims abstract description 6
- 238000011156 evaluation Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 230000000007 visual effect Effects 0.000 claims description 7
- 238000013480 data collection Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000003860 storage Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 230000006399 behavior Effects 0.000 claims description 4
- 238000010219 correlation analysis Methods 0.000 claims description 4
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000012217 deletion Methods 0.000 claims description 3
- 230000037430 deletion Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000007418 data mining Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 5
- 238000004590 computer program Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013479 data entry Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 229920006395 saturated elastomer Polymers 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
Classifications
-
- 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
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of data analysis and mining, and discloses a cloud computer customer loss early warning method and a system in the cloud computing industry, wherein the method comprises the following steps: s1: performing business analysis and data analysis to determine factors of customer loss; s2: collecting customer information based on factors of customer churn; s3: preprocessing customer information; s4: building a training set and a testing set of a customer loss prediction model; s5: establishing a customer loss prediction model, initializing parameters, training, evaluating, optimizing, storing and deploying the customer loss prediction model with optimal application; s6: inputting the client information into a client loss prediction model, and carrying out saving marketing on the predicted result which is the user of the loss client. The invention can predict and analyze the customer loss of the product, construct a stable customer loss prediction model and realize accurate early warning of the customer loss of the product.
Description
Technical Field
The invention belongs to the technical field of data analysis and mining, and mainly relates to a cloud computer customer loss early warning method and system in the cloud computing industry.
Background
Market competition is increasing and the cost of drawing new is increasing, for a mature product and saturated market, the cost of acquiring a new user is several times that of retaining an old user, and the reduction in churn rate means an increase in revenue.
With the gradual perfection of marketing systems, management of basic marketing and refined operation and management changes bring two main problems to business operations:
1. cloud merchants do not classify the loss of cloud computer products and perform customer image drawing, historical data fall asleep in a database, and business data lack deep analysis;
2. when the cloud merchant divides clients of the cloud computer product, the clients are simply divided according to simple rules such as resources, types and the like, then the value generated by the clients of the cloud computer product is simply divided into high, medium and low layers, the generated classification result is rough, certain scientific basis is lacked, pertinence and variability cannot be achieved when the cloud merchant finely plans the clients of the cloud computer product, and poor marketing effect is caused.
In view of this, there is still a need in the cloud business field for a sophisticated method and system for predicting and analyzing customer loss of products, and constructing a stable customer loss prediction model to realize accurate early warning of customer loss of products.
The patent with the application publication number of CN113379452A discloses a mobile banking customer loss early warning method and system, wherein the method comprises the following steps: acquiring login data of a mobile phone bank of a client, and defining the client which does not login the mobile phone bank in a first preset number of days as a lost client; acquiring user data of the loss clients, and performing feature extraction on the user data to obtain feature data; dividing the characteristic data into training data and verification data according to a preset proportion; training the mobile banking customer loss early warning model through the training data, and ending training when the training error value meets a preset value; and inputting the verification data into a trained mobile banking customer loss early warning model to verify the mobile banking customer loss early warning effect. The user data obtained by the method is comprehensive, and the accuracy of early warning prediction is improved.
The patent with the application publication number of CN115526652A discloses a customer loss early warning method and system based on machine learning, comprising the following steps: acquiring a data set; preprocessing data in a data set; constructing a customer loss early warning model, and training through the preprocessed data to obtain a trained customer loss early warning model; and inputting the data to be detected into a trained customer loss early warning model after preprocessing to obtain a customer loss early warning result. According to the invention, a large number of non-standard customer store information data of an enterprise are preprocessed, a decision tree, an SVM (support vector machine), a random forest and a GA-CART algorithm are adopted to construct a loss early warning model, and visualized manufacturing operation software is performed based on Python-Tlater, so that enterprise personnel are effectively assisted in carrying out customer loss early warning, and rescue measures are formulated.
The problems presented in the background art exist in the above patents: the lack of deep analysis on historical data leads to rough classification results, and pertinence and variability cannot be achieved when cloud merchants finely apply to clients of products, so that poor marketing effects are caused.
The information disclosed in this background section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person of ordinary skill in the art.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a cloud computer customer loss early warning method and a cloud computer customer loss early warning system in the cloud computing industry, which are used for carrying out predictive analysis on customer loss of products, constructing a stable customer loss prediction model and realizing accurate early warning on the customer loss of the products.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the invention provides a cloud computing industry cloud computer customer loss early warning method, which comprises the following steps:
s1: performing business analysis and data analysis to determine factors of customer loss;
s2: collecting client information based on factors of client loss and judging the client type;
s3: preprocessing customer information;
s4: building a training set and a testing set of a customer loss prediction model;
s5: establishing a customer loss prediction model, initializing parameters, training, evaluating, optimizing, storing and deploying the customer loss prediction model with optimal application;
s6: inputting the client information into a client loss prediction model, and carrying out saving marketing on the predicted result which is the user of the loss client.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the loss factors comprise active loss factors and passive loss factors, wherein the active loss factors comprise no use requirement, poor product experience, high cost and low price attraction of competing products; passive churn factors include offending use, arrears, expiration.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the customer information includes customer base information, order product information, product resource information, channel and order information, login behavior information, income level information, and transaction record information.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the client types include a churn client and a non-churn client; the method for determining that the client is a churn is as follows: if the cloud computer resource of the client is used for a period of time and the payment is not made within a specified date, the client is a loss client; otherwise, the customer is a non-attrition customer.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the preprocessing comprises null value deletion, abnormal value replacement and normalization processing; wherein, the empty value in the customer information is found and deleted through control judgment; outliers in the customer information are found by box graphs and replaced with median.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the training set and the test set are constructed by the following steps:
s41: collecting customer information to construct samples, each sample containing a customer type of a customer and customer information of approximately 30 days;
s42: adopting an SMOTE algorithm to expand the sample number of which the client type is the loss client;
s43: all samples were divided into training and testing sets, where the training and testing sets contained a 7:3 ratio of the number of samples.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the method for expanding the sample number of the client type which is the loss client comprises the following steps:
s4201: randomly extracting a sample of which the client type is a loss client, and marking the sample as A;
s4202: calculating k neighbors of A in all samples of which the client type is a churn client based on the Euclidean distance;
s4203: randomly selecting n samples from the K neighbors, and respectively carrying out random linear interpolation on the n samples and A to generate n new samples with the client type of loss clients;
s4204: steps S4201-S4203 are repeated until the customer type is the sample number balance of the churn customers and the non-churn customers.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the customer loss prediction model is built based on a LightGBM algorithm, input is customer information, and output is a predicted customer type; the training process is as follows:
s51: calculating the importance of each piece of client information to the client type through correlation analysis, and sorting the importance of each piece of client information;
s52: randomly extracting N samples from the training set with a put-back place, and inputting the client information of each sample into the model;
s53: calculating information gains of M pieces of client information with highest importance ranking;
s54: selecting the customer information with the largest information gain, dividing the sample into different subsets, and removing the selected customer information from the customer information of each sample;
s55: repeating the steps S53-S54 until the decision tree reaches the maximum depth.
As a preferable scheme of the cloud computer customer loss early warning method in the computing industry, the invention comprises the following steps: the indexes of the model evaluation comprise hit rate, coverage rate, lifting ratio and F1 value;
parameters related to model tuning comprise a learning rate, iteration times, a sub-sampling number N, a column sampling number M, a leaf node number and a maximum depth; after each training and evaluation and tuning, saving model parameters and evaluation results; and selecting a model deployment application with an optimal evaluation result.
The invention provides a cloud computing industry cloud computer customer loss early warning system, which comprises a data collection module, a data preprocessing module, a characteristic engineering module, a model training module, a model saving and deploying module, a saving marketing module and a visual report module, wherein:
the data collection module is used for collecting client information;
the data preprocessing module is used for preprocessing the client information;
the feature engineering module is used for carrying out feature extraction and conversion on the preprocessed customer information to generate a sample for training a customer loss prediction model;
the model training module is used for establishing a customer loss prediction model, initializing parameters, training, evaluating and optimizing parameters;
the model saving and deploying module is used for saving and deploying a client loss prediction model which is trained by the application;
the rescue marketing module is used for providing rescue marketing strategies for potential loss clients;
the visual report module is used for visually displaying the prediction result of the customer loss prediction model.
Compared with the prior art, the invention has the following beneficial effects:
in the method, in the construction of a cloud computer customer loss prediction model, the cloud computer customer loss relation is acquired through multiple dimensions of data such as user basic information, charging information, product resource information, customer information, login behavior information and the like, a reasonable cloud computer loss early warning and identifying scheme is designed, cloud computer customers needing to be focused are identified by combining algorithms, different propaganda and marketing strategies are formulated according to the requirements of different cloud computer customers and contributions to enterprises, the sales cost of the enterprises is reduced, the operating efficiency of cloud products and the competitiveness of the enterprises are improved, and the problems of incomplete basis based on traditional model identification, poor stability and low accuracy of the obtained results are solved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:
FIG. 1 is a flow chart of a cloud computing industry cloud computer customer churning early warning method provided by the invention;
fig. 2 is a schematic diagram of a result of customer information importance ranking provided by the present invention.
Detailed Description
The following detailed description of the present invention is made with reference to the accompanying drawings and specific embodiments, and it is to be understood that the specific features of the embodiments and the embodiments of the present invention are detailed description of the technical solutions of the present invention, and not limited to the technical solutions of the present invention, and that the embodiments and the technical features of the embodiments of the present invention may be combined with each other without conflict.
Example 1
The embodiment describes a cloud computing industry cloud computer customer loss early warning method, referring to fig. 1, the method comprises the following steps:
s1: performing business analysis and data analysis to determine factors of customer loss;
the loss factors comprise active loss factors and passive loss factors, wherein the active loss factors comprise no use requirement, poor product experience, high cost and low price attraction of competing products; passive churn factors include offending use, arrears, expiration.
S2: collecting client information based on factors of client loss and judging the client type;
the customer information includes customer base information, order product information, product resource information, channel and order information, login behavior information, income level information, and transaction record information.
The client types include a churn client and a non-churn client; the method for determining that the client is a churn is as follows: if the cloud computer resource of the client is used for a period of time and the payment is not made within a specified date, the client is a loss client; otherwise, the customer is a non-attrition customer.
Wherein, the definition details of the client information contained in the client basic information are shown in table 1:
TABLE 1 customer base definition details
S3: preprocessing customer information;
there may be outliers or outliers in the dataset that are not coincident with other data points. These outliers may be due to erroneous measurement, recording or data entry errors. In the building of customer churn prediction models, these outliers may negatively impact the performance and generalization ability of the model, and thus need to be handled.
The preprocessing comprises null value deletion, abnormal value replacement and normalization processing; wherein, the empty value in the customer information is found and deleted through control judgment; outliers in the customer information are found by box graphs and replaced with median.
S4: building a training set and a testing set of a customer loss prediction model;
the training set and the test set are constructed by the following steps:
s41: collecting customer information to construct samples, each sample containing a customer type of a customer and customer information of approximately 30 days;
s42: adopting an SMOTE algorithm to expand the sample number of which the client type is the loss client; the method comprises the following steps: the method for expanding the sample number of the client type which is the loss client comprises the following steps:
s4201: randomly extracting a sample of which the client type is a loss client, and marking the sample as A;
s4202: calculating k neighbors of A in all samples of which the client type is a churn client based on the Euclidean distance;
s4203: randomly selecting n samples from the K neighbors, and respectively carrying out random linear interpolation on the n samples and A to generate n new samples with the client type of loss clients;
s4204: steps S4201-S4203 are repeated until the customer type is the sample number balance of the churn customers and the non-churn customers.
The samples of primary interest to the customer churn prediction model are those of the churn customer type, and their number is relatively small. Therefore, in order for the model to better predict customer types, a sample of customer types that are attrition customers is augmented.
S43: all samples were divided into training and testing sets, where the training and testing sets contained a 7:3 ratio of the number of samples.
S5: establishing a customer loss prediction model, initializing parameters, training, evaluating, optimizing, storing and deploying the customer loss prediction model with optimal application;
the customer loss prediction model is built based on a LightGBM algorithm, input is customer information, and output is a predicted customer type; the training process is as follows:
s51: calculating the importance of each piece of client information to the client type through correlation analysis, and sorting the importance of each piece of client information; the partial results of the importance ranking are shown in figure 2.
Through correlation analysis, customer information with higher correlation with the customer type is selected, and features with lower correlation are removed, so that the customer information with the most influence is selected and used for training a model, and the accuracy and generalization capability of the model are improved.
S52: randomly extracting N samples from the training set with a put-back place, and inputting the client information of each sample into the model;
s53: calculating information gains of M pieces of client information with highest importance ranking;
s54: selecting the customer information with the largest information gain, dividing the sample into different subsets, and removing the selected customer information from the customer information of each sample;
s55: repeating the steps S53-S54 until the decision tree reaches the maximum depth.
The indexes of the model evaluation comprise hit rate, coverage rate, lifting ratio and F1 value;
among the above indexes, the hit rate is emphasized. The purpose of identifying potential customers who pass is to market them, so we can expand the scope of marketing to cover more potential customers who are lost, and can include a higher percentage of non-potential customers who are lost, if cost allows.
Parameters related to model tuning comprise a learning rate, iteration times, a sub-sampling number N, a column sampling number M, a leaf node number and a maximum depth; after each training and evaluation and tuning, saving model parameters and evaluation results; and selecting a model deployment application with an optimal evaluation result.
S6: inputting the client information into a client loss prediction model, and carrying out saving marketing on the predicted result which is the user of the loss client.
And constructing a client list of potential loss clients according to the model prediction result. The feasible product functions are deduced through user investigation and invite users in the list to try out, and the following aspects are specifically covered:
the customized service such as upgrading package, weighting benefit package, etc. is pushed out, on one hand, the association degree with other users is enhanced, and on the other hand, personalized service is provided for specific users
The service of the package month such as free resource upgrading, long time length and the like is promised for the user so as to improve the viscosity of the user; aiming at value-added services such as online security, online backup, equipment protection, technical support and the like, the popularization and introduction of users, such as first month/half year free experience, should be emphasized.
Aiming at a single-month contract client, pushing out a year contract payment discount activity, converting the month contract client into a year contract client, and improving the online time of the client so as to achieve higher client retention; for a customer paying by adopting electronic checks, coupons for pushing other payment modes are suggested to be directed, and the customer is guided to change the payment modes.
Example 2
This embodiment is a second embodiment of the present invention; the embodiment of the invention, which is the same as that of embodiment 1, introduces a cloud computing industry cloud computer customer loss early warning system, which comprises: the system comprises a data collection module, a data preprocessing module, a characteristic engineering module, a model training module, a model storage and deployment module, a saving marketing module and a visual report module, wherein:
the data collection module is used to collect customer information from a plurality of data sources.
The data preprocessing module is used for preprocessing the collected customer information, such as cleaning, deduplication, missing value processing, abnormal value processing and the like, so as to ensure the quality and the integrity of the data.
And the feature engineering module performs feature extraction and conversion on the preprocessed customer information according to the service demand and the feature selection method to generate a sample for modeling input.
The model training module adopts a LightGBM algorithm to establish a customer loss prediction model, initialize parameters, train, evaluate and tune.
The model saving and deploying module is used for saving the trained customer churn prediction model and deploying the model into an available application program.
The saving marketing module is used for providing corresponding saving marketing strategies and measures, such as personalized recommendation, preferential activities and the like, for potential loss clients according to the prediction result of the client loss prediction model.
The visual report module is used for displaying the prediction result of the customer loss prediction model in a visual mode to generate a report and a chart, so that business personnel can quickly know the customer loss condition and formulate a saving marketing strategy.
The specific functions of the above modules are implemented by referring to the relevant content in the cloud computing client loss early warning method in the cloud computing industry in embodiment 1, and will not be described in detail.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present invention and the scope of the claims, which are all within the protection of the present invention.
Claims (10)
1. A cloud computing industry cloud computer customer loss early warning method is characterized in that: the method comprises the following steps:
s1: performing business analysis and data analysis to determine factors of customer loss;
s2: collecting client information based on factors of client loss and judging the client type;
s3: preprocessing customer information;
s4: building a training set and a testing set of a customer loss prediction model;
s5: establishing a customer loss prediction model, initializing parameters, training, evaluating, optimizing, storing and deploying the customer loss prediction model with optimal application;
s6: inputting the client information into a client loss prediction model, and carrying out saving marketing on the predicted result which is the user of the loss client.
2. The cloud computing industry cloud computer customer churn early warning method according to claim 1, wherein: the loss factors comprise active loss factors and passive loss factors, wherein the active loss factors comprise no use requirement, poor product experience, high cost and low price attraction of competing products; passive churn factors include offending use, arrears, expiration.
3. The cloud computing industry cloud computer customer churn early warning method according to claim 2, wherein: the customer information includes customer base information, order product information, product resource information, channel and order information, login behavior information, income level information, and transaction record information.
4. The cloud computing industry cloud computer customer churn early warning method according to claim 3, wherein: the client types include a churn client and a non-churn client; the method for determining that the client is a churn is as follows: if the cloud computer resource of the client is used for a period of time and the payment is not made within a specified date, the client is a loss client; otherwise, the customer is a non-attrition customer.
5. The cloud computing industry cloud computer customer churn early warning method according to claim 4, wherein: the preprocessing comprises null value deletion, abnormal value replacement and normalization processing; wherein, the empty value in the customer information is found and deleted through control judgment; outliers in the customer information are found by box graphs and replaced with median.
6. The cloud computing industry cloud computer customer churn early warning method according to claim 5, wherein: the training set and the test set are constructed by the following steps:
s41: collecting customer information to construct samples, each sample containing a customer type of a customer and customer information of approximately 30 days;
s42: adopting an SMOTE algorithm to expand the sample number of which the client type is the loss client;
s43: all samples were divided into training and testing sets, where the training and testing sets contained a 7:3 ratio of the number of samples.
7. The cloud computing industry cloud computer customer churn early warning method according to claim 6, wherein: the method for expanding the sample number of the client type which is the loss client comprises the following steps:
s4201: randomly extracting a sample of which the client type is a loss client, and marking the sample as A;
s4202: calculating k neighbors of A in all samples of which the client type is a churn client based on the Euclidean distance;
s4203: randomly selecting n samples from the K neighbors, and respectively carrying out random linear interpolation on the n samples and A to generate n new samples with the client type of loss clients;
s4204: steps S4201-S4203 are repeated until the customer type is the sample number balance of the churn customers and the non-churn customers.
8. The cloud computing industry cloud computer customer churn early warning method according to claim 7, wherein: the customer loss prediction model is built based on a LightGBM algorithm, input is customer information, and output is a predicted customer type; the training process is as follows:
s51: calculating the importance of each piece of client information to the client type through correlation analysis, and sorting the importance of each piece of client information;
s52: randomly extracting N samples from the training set with a put-back place, and inputting the client information of each sample into the model;
s53: calculating information gains of M pieces of client information with highest importance ranking;
s54: selecting the customer information with the largest information gain, dividing the sample into different subsets, and removing the selected customer information from the customer information of each sample;
s55: repeating the steps S53-S54 until the decision tree reaches the maximum depth.
9. The cloud computing industry cloud computer customer churn early warning method according to claim 8, wherein: the indexes of the model evaluation comprise hit rate, coverage rate, lifting ratio and F1 value;
parameters related to model tuning comprise a learning rate, iteration times, a sub-sampling number N, a column sampling number M, a leaf node number and a maximum depth; after each training and evaluation and tuning, saving model parameters and evaluation results; and selecting a model deployment application with an optimal evaluation result.
10. A cloud computing industry cloud computer customer loss early warning system, implemented based on the cloud computing industry cloud computer customer loss early warning method of any one of claims 1-9, characterized in that: the system comprises a data collection module, a data preprocessing module, a characteristic engineering module, a model training module, a model storage and deployment module, a saving marketing module and a visual report module, wherein:
the data collection module is used for collecting client information;
the data preprocessing module is used for preprocessing the client information;
the feature engineering module is used for carrying out feature extraction and conversion on the preprocessed customer information to generate a sample for training a customer loss prediction model;
the model training module is used for establishing a customer loss prediction model, initializing parameters, training, evaluating and optimizing parameters;
the model saving and deploying module is used for saving and deploying a client loss prediction model which is trained by the application;
the rescue marketing module is used for providing rescue marketing strategies for potential loss clients;
the visual report module is used for visually displaying the prediction result of the customer loss prediction model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311703144.9A CN117788043A (en) | 2023-12-12 | 2023-12-12 | Cloud computing industry cloud computer customer loss early warning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311703144.9A CN117788043A (en) | 2023-12-12 | 2023-12-12 | Cloud computing industry cloud computer customer loss early warning method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117788043A true CN117788043A (en) | 2024-03-29 |
Family
ID=90393677
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311703144.9A Pending CN117788043A (en) | 2023-12-12 | 2023-12-12 | Cloud computing industry cloud computer customer loss early warning method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117788043A (en) |
-
2023
- 2023-12-12 CN CN202311703144.9A patent/CN117788043A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110837931B (en) | Customer churn prediction method, device and storage medium | |
| CN108564286B (en) | Artificial intelligent financial wind-control credit assessment method and system based on big data credit investigation | |
| CN111291816B (en) | Method and device for carrying out feature processing aiming at user classification model | |
| CN109583966B (en) | High-value customer identification method, system, equipment and storage medium | |
| CN107993143A (en) | A kind of Credit Risk Assessment method and system | |
| CN110263827B (en) | Abnormal transaction detection method and device based on transaction rule identification | |
| CN112700325A (en) | Method for predicting online credit return customers based on Stacking ensemble learning | |
| CN114048436A (en) | Construction method and construction device for forecasting enterprise financial data model | |
| CN112419030A (en) | Method, system and equipment for evaluating financial fraud risk | |
| CN110674100B (en) | User demand prediction method and framework based on full-channel operation data | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| CN112990989A (en) | Value prediction model input data generation method, device, equipment and medium | |
| CN116821759A (en) | Identification prediction method and device for category labels, processor and electronic equipment | |
| CN116468273A (en) | Customer risk identification method and device | |
| CN115563176A (en) | Electronic commerce data processing system and method | |
| CN117788043A (en) | Cloud computing industry cloud computer customer loss early warning method and system | |
| CN114626940A (en) | Data analysis method and device and electronic equipment | |
| CN112926989A (en) | Financial transaction risk assessment method and device based on multi-view ensemble learning | |
| CN112950392A (en) | Information display method, posterior information determination method and device and related equipment | |
| Benavente-Peces et al. | Applied Data Analytics | |
| CN117436918A (en) | Method and system for constructing mobile banking client liveness layered model | |
| CN117649293A (en) | Asset retention promotion method and system for bank-oriented issuing clients | |
| Nasution et al. | Credit Risk Detection in Peer-to-Peer Lending Using CatBoost | |
| CN117557297A (en) | Method and system for predicting willingness of cross marketing users | |
| CN117670350A (en) | Transaction anti-fraud early warning method and device based on multi-model integration |
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 |