Telecommunication enterprise customer loss prediction method, system and equipment based on random forest and medium
By adopting a random forest-based method in the prediction of customer churn of telecom enterprises, an integrated model of multiple decision trees is solved, and the problems of low prediction accuracy and overfitting in the existing technology are achieved, and more efficient and accurate customer churn prediction is achieved, helping telecom enterprises improve user satisfaction and operational cost control.
Patent Information
- Application Number
- CN202510168715.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-17
- Publication Date
- 2025-06-13
AI Technical Summary
The prediction accuracy rate of the existing technology is not high in the prediction of customer churn prediction of telecommunications enterprises, and there are problems such as overfitting, making it difficult to formulate effective personalized retention measures, which affects users' service perception and satisfaction and operator image.
The random forest-based customer churn prediction method of telecommunications enterprises is adopted. By collecting customer feature information and behavior information, data preprocessing and feature selection are carried out, a random forest model is constructed, and the prediction accuracy and stability are improved by using Bootstrap sampling and integrated learning of multiple decision trees.
It significantly improves the accuracy and stability of customer churn forecasts, helps telecom companies to discover customers with churn tendencies in advance, formulate personalized retention measures, improves users' service perception and satisfaction, and operator image, and reduces operating costs.
Smart Images
Figure CN120146258A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of telecommunications operation, and specifically relates to a method, system, device and medium for predicting customer churn of telecommunications enterprises based on random forest. Background Technique
[0002] In recent years, with the rapid economic development and the improvement of information infrastructure construction, the scale of operator customers has become larger and larger. Although having the world's largest telecommunications user market, the telecommunications user market has approached saturation, and the competition among operators in the existing user group has become very fierce. In order to promote the improvement of the service quality of operators, "number portability" has become popular. The number portability service allows users to choose a telecommunications operator with better signal, higher package cost performance and better service quality without changing the number. According to the data statistics of a certain telecommunications enterprise, the proportion of number portability users in the total number of mobile lost users is getting higher and higher. This trend indicates that the implementation of the number portability policy has subverted the market pattern of telecommunications operators, brought greater challenges to the retention of existing users, and made user churn more frequent. Therefore, putting the focus of work on retaining old users and formulating corresponding retention measures before user churn has become an important task that operators need to face.
[0003] In the fierce competition in the telecommunications industry, customer churn is one of the important challenges faced by telecommunications enterprises. Traditional customer churn prediction methods mostly rely on statistical models or single machine learning algorithms, and there are problems such as low prediction accuracy and overfitting.
[0004] Therefore, how to accurately and efficiently predict the tendency of telecommunications customer churn, formulate personalized retention measures, and then improve the service perception and satisfaction of users as well as the image of operators is a technical problem that needs to be solved urgently at present. Summary of the Invention
[0005] The technical task of the present invention is to provide a method, system, device and medium for predicting customer churn of telecommunications enterprises based on random forest to solve the problem of how to accurately and efficiently predict the tendency of telecommunications customer churn, formulate personalized retention measures, and then improve the service perception and satisfaction of users as well as the image of operators.
[0006] The technical task of the present invention is realized in the following way. A method for predicting customer churn of telecommunications enterprises based on random forest is as follows:
[0007] Data collection: Collect customer feature information and customer behavior information labels to construct a data set; among them, the customer feature information and customer behavior information labels include user call information, user Internet behavior data, user SMS information, value-added service data, and porting-related information;
[0008] Data preprocessing: Perform preprocessing operations on customer feature data and behavioral data, including handling missing values, converting categorical variables (such as converting strings to numerical types), and standardizing numerical variables (such as using one-hot encoding to handle categorical variables like gender and contract type), to obtain the feature data of the customer to be predicted;
[0009] Feature selection: Use the random forest algorithm to evaluate the importance of each customer feature during training, and select the features that have a significant impact on the prediction result as input variables;
[0010] Construct a random forest model: Adopt the Bootstrap sampling method to randomly draw multiple subsets from the original training set with replacement, and each subset is used to train a decision tree; when constructing each decision tree, randomly select some features for node splitting to increase the diversity of the model, and train multiple decision trees to form a random forest model;
[0011] Model training and tuning: Use the training set to train the random forest model, optimize the model performance by adjusting parameters related to the number of trees, the maximum number of features, and the maximum depth, and use the Out-of-Bag samples to evaluate the model performance without an additional validation set;
[0012] Predict customer churn: Input the feature data of the customer to be predicted into the trained random forest model, obtain the prediction result through the majority voting method (for classification problems) or the averaging method (for regression problems), and judge the customer churn tendency based on the prediction result, providing a basis for the enterprise to formulate targeted customer retention strategies.
[0013] Preferably, during the data collection process, read the data from the CSV file through the read.csv() function and use the str() function to check the structure of the data.
[0014] Preferably, during the data preprocessing process, complete the preprocessing operations through functions in R. Specifically, when handling missing values, use the na.omit() function to delete the rows containing missing values, or use the mean or median method for filling.
[0015] Preferably, the feature selection is as follows:
[0016] Eliminate the skewness of numerical prediction variables;
[0017] Standardize all digital prediction variables;
[0018] Create dummy variables for all nominal prediction variables.
[0019] Preferably, in R, train the random forest model through the randomForest() function.
[0020] Preferably, the performance of the random forest model is evaluated by the confusion matrix and ROC curve methods.
[0021] A customer churn prediction system for telecommunications enterprises based on random forest, the system includes:
[0022] A data collection module, used to collect customer feature information and customer behavior information labels to construct a data set; among them, the customer feature information and customer behavior information labels include user call information, user Internet behavior data, user SMS information, value-added service data, and porting-related information;
[0023] A data preprocessing module, used to perform preprocessing operations on customer feature data and behavior data, such as handling missing values, converting categorical variables (such as converting strings to numerical types), and standardizing numerical variables (such as using one-hot encoding to handle categorical variables such as gender and contract type), to obtain the feature data of the customers to be predicted;
[0024] A feature selection module, used to evaluate the importance of each customer feature during the training process using the random forest algorithm, and select the features that have a significant impact on the prediction result as input variables;
[0025] A random forest model construction module, used to adopt the Bootstrap sampling method to randomly draw multiple subsets from the original training set with replacement, and each subset is used to train a decision tree; when constructing each decision tree, randomly select some features for node splitting to increase the diversity of the model, and train multiple decision trees to form a random forest model;
[0026] A training and tuning module, used to train the random forest model using the training set, optimize the model performance by adjusting parameters related to the number of trees, the maximum number of features, and the maximum depth, and use the Out-of-Bag samples to evaluate the model performance without an additional validation set;
[0027] A customer churn prediction module, used to input the feature data of the customers to be predicted into the trained random forest model, obtain the prediction result through the majority voting method (for classification problems) or the averaging method (for regression problems), and judge the customer churn tendency according to the prediction result, providing a basis for the enterprise to formulate targeted customer retention strategies.
[0028] Preferably, the data collection module reads data from a CSV file through the read.csv() function and uses the str() function to check the structure of the data;
[0029] The data preprocessing module completes the preprocessing operations through functions in R. Specifically, when handling missing values, the na.omit() function is used to delete the rows containing missing values or the mean and median methods are used for filling;
[0030] The working process of the feature selection module includes eliminating the skewness of numerical predictive variables, normalizing all numerical predictive variables, and creating dummy variables for all nominal predictive variables.
[0031] An electronic device, comprising: a memory and at least one processor;
[0032] Wherein, a computer program is stored on the memory;
[0033] The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the above-mentioned customer churn prediction method for telecommunications enterprises based on random forest.
[0034] A computer-readable storage medium stores a computer program, and the computer program can be executed by a processor to implement the above-mentioned customer churn prediction method for telecommunications enterprises based on random forest.
[0035] The customer churn prediction method, system, device and medium based on random forest of the present invention have the following advantages:
[0036] (1) The present invention is based on the random forest algorithm, using the multi-dimensional customer data collected as input, and using the core algorithm steps of the random forest algorithm such as constructing decision trees, feature selection, decision tree training, and prediction to predict the churn trend of telecommunications users;
[0037] (2) The present invention can help maintain the stability of the income of the existing customers of telecommunications enterprises, thereby reducing the operating costs and increasing the operating profits of the enterprises;
[0038] (3) The present invention improves the prediction accuracy, stability and generalization ability of the overall model by constructing multiple decision trees and summarizing their prediction results. Random forest belongs to an implementation of the "bagging" (Bootstrap Aggregating) method, which combines the powerful classification ability of decision trees and the advantages of ensemble learning;
[0039] (4) The present invention improves the accuracy, stability and generalization ability of customer churn prediction by constructing multiple decision trees and performing ensemble learning, and has the advantages of high accuracy, strong robustness and easy use;
[0040] (5) The present invention improves the prediction accuracy: by integrating the prediction results of multiple decision trees, the random forest algorithm significantly improves the accuracy of customer churn prediction;
[0041] (6) The present invention enhances the model stability: the random forest algorithm is insensitive to the noise and outliers of the training data and has strong robustness;
[0042] (7) The present invention is easy to parallelize: Each decision tree can be trained independently and is naturally suitable for parallel computing, which improves the model training efficiency.
[0043] (8) The present invention provides feature importance evaluation: Evaluating the importance of each feature during the training process helps with feature selection and model interpretation.
[0044] (9) The present invention reduces the operating costs of telecommunications enterprises and provides the stability of the income of existing customers: It can help telecommunications enterprises discover customers with a tendency to churn in advance, and then formulate personalized retention measures to enhance the service perception and satisfaction of users and the image of the operator. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] The present invention will be further described below in conjunction with the drawings.
[0046] Att Figure 1 is a flowchart of a method for predicting customer churn in a telecommunications enterprise based on random forest;
[0047] Att Figure 2 is an ROC curve graph using the random forest algorithm. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0048] The method, system, device, and medium for predicting customer churn in a telecommunications enterprise based on random forest of the present invention will be described in detail below with reference to the drawings of the specification and specific embodiments.
[0049] Example 1:
[0050] As shown in Att Figure 1 This embodiment provides a method for predicting customer churn in a telecommunications enterprise based on random forest, and the method is as follows:
[0051] S1. Data collection: Collect customer feature information and customer behavior information labels to construct a data set; among them, the customer feature information and customer behavior information labels include user call information, user Internet behavior data, user text message information, value-added service data, and porting-related information;
[0052] S2. Data preprocessing: Perform preprocessing operations on customer feature data and behavior data, such as handling missing values, converting categorical variables (such as converting strings to numerical types), and standardizing numerical variables (such as using one-hot encoding to handle categorical variables such as gender and contract type) to obtain the feature data of the customers to be predicted;
[0053] S3. Feature selection: Use the random forest algorithm to evaluate the importance of each customer feature during the training process, and select the features that have a significant impact on the prediction result as input variables;
[0054] S4. Construct a random forest model: Using the Bootstrap sampling method, randomly draw multiple subsets from the original training set with replacement, and each subset is used to train a decision tree; when constructing each decision tree, randomly select some features for node splitting to increase the diversity of the model, and train multiple decision trees to form a random forest model;
[0055] S5. Model training and tuning: Use the training set to train the random forest model, optimize the model performance by adjusting parameters related to the number of trees, the maximum number of features, and the maximum depth, and use the Out-of-Bag samples to evaluate the model performance without an additional validation set;
[0056] S6. Predict customer churn: Input the feature data of the customers to be predicted into the trained random forest model, obtain the prediction results through the majority voting method (for classification problems) or the averaging method (for regression problems), and judge the customer churn tendency based on the prediction results to provide a basis for the enterprise to formulate targeted customer retention strategies.
[0057] In the data collection process of step S1 of this embodiment, data is read from the CSV file through the read.csv() function, and the str() function is used to check the structure of the data.
[0058] In the data preprocessing process of step S2 of this embodiment, the preprocessing operations are completed through functions in R. Specifically, when dealing with missing values, the na.omit() function is used to delete the rows containing missing values, or the mean and median methods are used for filling.
[0059] The feature selection in step S3 of this embodiment is as follows:
[0060] S301. Eliminate the skewness of numerical predictive variables;
[0061] S302. Standardize all numerical predictive variables;
[0062] S303. Create dummy variables for all nominal predictive variables.
[0063] In this embodiment, in R, the random forest model is trained through the randomForest() function.
[0064] In this embodiment, the performance of the random forest model is evaluated through the confusion matrix and the ROC curve method, as shown in the appendix Figure 2 as follows.
[0065] The key code is as follows:
[0066] import pandas as pd
[0067] from sklearn.model_selection import train_test_split
[0068] from sklearn.ensemble import RandomForestClassifier
[0069] from sklearn.metrics import accuracy_score, classification_report
[0070] from sklearn.preprocessing import OneHotEncoder
[0071] from sklearn.compose import ColumnTransformer
[0072] from sklearn.pipeline import Pipeline
[0073] # Load data
[0074] data = pd.read_csv('customer_data.csv')
[0075] # Assume the last column is the target variable (churn, 1 means churn, 0 means not churn)
[0076] X = data.iloc[:, :-1]
[0077] y = data.iloc[:, -1]
[0078] # Split the training set and the test set
[0079] X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)
[0080] # Define numerical features and categorical features
[0081] It should be noted that there seems to be a misspelling in the original text where "train_test_spl it" should be "train_test_split". This has been corrected in the translation. Also, it's assumed that "pd" is a pre-defined variable for the pandas library which is not shown in the provided code but is common in Python data analysis scenarios.numeric_features = X_train.select_dtypes(include=['int64', 'float64']).columns categorical_features = X_train.select_dtypes(include=['object']).columns
[0082] # Create preprocessing steps
[0083] numeric_transformer = Pipeline(steps=[]) # For numeric features, there are no special preprocessing steps here
[0084] categorical_transformer = Pipeline(steps=
[0085] ('onehot', OneHotEncoder(handle_unknown='ignore')) )
[0087] # Combine the features
[0088] preprocessor = ColumnTransformer(
[0089] transformers =
[0090] ('num', numeric_transformer, numeric_features),
[0091] ('cat', categorical_transformer, categorical_features) )
[0093] # Create the model pipeline
[0094] pipeline = Pipeline(steps=
[0095] ('preprocessor', preprocessor),
[0096] ('classifier', RandomForestClassifier(n_estimators = 100, random_state = 42)) )
[0098] # Train the model
[0099] pipeline.fit(X_train, y_train)
[0100] # Predict the test set
[0101] y_pred = pipeline.predict(X_test)
[0102] # Evaluate the model
[0103] accuracy = accuracy_score(y_test, y_pred)
[0104] print(f'Accuracy: {accuracy}')
[0105] print(classification_report(y_test, y_pred))
[0106] Example 2:
[0107] This example provides a customer churn prediction system for telecommunications enterprises based on random forest. The system includes:
[0108] A data collection module for collecting customer feature information and customer behavior information labels to construct a dataset. Among them, the customer feature information and customer behavior information labels include user call information, user Internet behavior data, user SMS information, value-added service data, and porting-related information;
[0109] A data preprocessing module for performing preprocessing operations on customer feature data and behavior data, such as handling missing values, converting categorical variables (such as converting strings to numerical types), and standardizing numerical variables (such as using one-hot encoding to handle categorical variables such as gender and contract type) to obtain the feature data of the customers to be predicted;
[0110] A feature selection module for using the random forest algorithm to evaluate the importance of each customer feature during training and selecting the features that have a significant impact on the prediction result as input variables;
[0111] A random forest model construction module for using the Bootstrap sampling method to randomly draw multiple subsets from the original training set with replacement, and each subset is used to train a decision tree. When constructing each decision tree, randomly select some features for node splitting to increase the diversity of the model, and train multiple decision trees to form a random forest model;
[0112] A training and tuning module for training a random forest model using a training set, optimizing the model performance by adjusting parameters related to the number of trees, the maximum number of features, and the maximum depth, and evaluating the model performance using out-of-bag samples without the need for an additional validation set;
[0113] A customer churn prediction module for inputting the feature data of the customer to be predicted into the trained random forest model, obtaining the prediction result by the majority voting method (for classification problems) or the averaging method (for regression problems), and judging the customer churn tendency according to the prediction result, providing a basis for the enterprise to formulate targeted customer retention strategies.
[0114] In this embodiment, the data collection module reads data from a CSV file through the read.csv() function and checks the structure of the data using the str() function.
[0115] In this embodiment, the data preprocessing module completes the preprocessing operations through functions in R. Specifically, when dealing with missing values, the na.omit() function is used to delete the rows containing missing values or the mean and median methods are used for filling.
[0116] The working process of the feature selection module in this embodiment includes eliminating the skewness of numerical predictive variables, normalizing all numerical predictive variables, and creating dummy variables for all nominal predictive variables.
[0117] Embodiment 3:
[0118] This embodiment of the present invention also provides an electronic device, including: a memory and a processor;
[0119] Wherein, the memory stores computer execution instructions;
[0120] The processor executes the computer execution instructions stored in the memory, so that the processor executes the random forest-based telecom enterprise customer churn prediction method in any embodiment of the present invention.
[0121] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0122] The memory can be used to store computer programs and / or modules. By running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory, the processor can implement various functions of the electronic device. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory can also include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash memory card, at least one magnetic disk storage period, a flash memory device, or other volatile solid-state storage devices.
[0123] Embodiment 4:
[0124] This embodiment also provides a computer-readable storage medium, in which multiple instructions are stored. The instructions are loaded by the processor to enable the processor to execute the method for predicting customer churn of telecommunications enterprises based on random forest in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided. On this storage medium, software program codes for implementing the functions of any one of the above embodiments are stored, and the computer (or CPU or MPU) of the system or device reads and executes the program codes stored in the storage medium.
[0125] In this case, the program code read from the storage medium itself can implement the functions of any one of the above embodiments. Therefore, the program code and the storage medium storing the program code constitute a part of the present invention.
[0126] Embodiments of the storage medium for providing program codes include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Optionally, the program code can be downloaded from a server computer via a communication network.
[0127] In addition, it should be clear that not only can the actual operations be completed in part or in whole by executing the program code read by the computer, but also by the operating system etc. operating on the computer based on the instructions of the program code, so as to implement the functions of any one of the above embodiments.
[0128] In addition, it can be understood that the program code read out from the storage medium is written into the memory provided in the expansion board inserted into the computer or into the memory provided in the expansion unit connected to the computer, and then based on the instructions of the program code, the CPU etc. installed on the expansion board or the expansion unit are made to execute part or all of the actual operations, thereby implementing the functions of any one of the above embodiments.
[0129] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for predicting customer churn in telecommunication enterprises based on random forest, characterized in that: The method is as follows: Data collection: Collect customer feature information and customer behavior information labels to build a data set; customer feature information and customer behavior information labels include user call information, user Internet behavior data, user SMS information, value-added service data, and portability related information; Data preprocessing: Perform preprocessing operations on customer feature data and behavior data to process missing values, convert categorical variables, and standardize numerical variables to obtain feature data of the customer to be predicted; Feature selection: Use the random forest algorithm to evaluate the importance of each customer feature during the training process and select the features that have a significant impact on the prediction results as input variables; Constructing a random forest model: Using the Bootstrap sampling method, multiple subsets are randomly extracted from the original training set with replacement, and each subset is used to train a decision tree. When constructing each decision tree, some features are randomly selected for node splitting to increase the diversity of the model, and multiple decision trees are trained to form a random forest model. Model training and tuning: Use the training set to train the random forest model, optimize the model performance by adjusting the number of trees, maximum number of features, and maximum depth-related parameters, and use out-of-bag samples to evaluate the model performance; Predict customer churn: Input the characteristic data of the customer to be predicted into the trained random forest model, obtain the prediction results through majority voting or averaging, and judge the customer churn tendency based on the prediction results, providing a basis for the company to formulate targeted customer retention strategies.
2. The method for predicting customer churn of telecommunication enterprises based on random forest according to claim 1, characterized in that: During the data collection process, the data is read from the CSV file using the read.csv() function and the str() function is used to check the structure of the data.
3. The method for predicting customer churn of telecommunication enterprises based on random forest according to claim 1, characterized in that: During data preprocessing, preprocessing operations are completed through functions in R. Specifically, when dealing with missing values, the na.omit() function is used to delete rows containing missing values, or the mean or median method is used to fill them.
4. The method for predicting customer churn of telecommunication enterprises based on random forest according to claim 1, characterized in that: The feature selection is as follows: Eliminate skewness of numerical predictors; Standardize all numeric predictors; Create dummy variables for all nominal predictors.
5. The method for predicting customer churn of telecommunication enterprises based on random forest according to claim 1, characterized in that: In R, the random forest model is trained using the randomForest() function.
6. The method for predicting customer churn of a telecommunication enterprise based on random forest according to any one of claims 1 to 5, characterized in that: The performance of the random forest model was evaluated using confusion matrix and ROC curve methods.
7. A telecommunication enterprise customer churn prediction system based on random forest, characterized in that: The system includes: The data collection module is used to collect customer feature information and customer behavior information tags to construct a data set; wherein the customer feature information and customer behavior information tags include user call information, user online behavior data, user text message information, value-added service data, and portability related information; The data preprocessing module is used to process missing values, convert categorical variables, and standardize numerical variables on customer feature data and behavior data to obtain feature data of the customer to be predicted; The feature selection module is used to evaluate the importance of each customer feature during the training process using the random forest algorithm and select the customer features that have a significant impact on the prediction results as input variables; The random forest model building module is used to randomly extract multiple subsets with replacement from the original training set using the Bootstrap sampling method, and each subset is used to train a decision tree. When building each decision tree, some features are randomly selected for node splitting to increase the diversity of the model, and multiple decision trees are trained to form a random forest model. The training and tuning module is used to train the random forest model using the training set, optimize the model performance by adjusting the number of trees, the maximum number of features, and the maximum depth-related parameters, and evaluate the model performance using out-of-bag samples; The customer churn prediction module is used to input the characteristic data of the customer to be predicted into the trained random forest model, obtain the prediction results through majority voting method or averaging method, and judge the customer churn tendency based on the prediction results, providing a basis for enterprises to formulate targeted customer retention strategies.
8. The telecommunication enterprise customer churn prediction system based on random forest according to claim 7, characterized in that: The data collection module reads data from the CSV file using the read.csv() function and checks the structure of the data using the str() function; The data preprocessing module completes the preprocessing operation through the functions in R. Specifically, when dealing with missing values, the na.omit() function is used to delete rows with missing values or to fill them with the mean or median method; The feature selection module works by eliminating the skewness of the numerical predictors, standardizing all the numeric predictors, and creating dummy variables for all the nominal predictors.
9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the telecommunication enterprise customer churn prediction method based on random forest as described in any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed by a processor to implement the telecommunication enterprise customer churn prediction method based on random forest as described in any one of claims 1 to 6.