AI Machine Learning-Based Automatic Customer Churn Early Warning Method and Device

Abstract

The present invention discloses an automatic early warning method and device for customer churn based on AI machine learning, and belongs to the field of data processing systems for management. The present invention obtains the sales record data of customers in the enterprise, and at the same time obtains the customer's complaint situation and transaction satisfaction data for each transaction, obtains the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and imports them into the customer churn primary judgment strategy to make a primary judgment on customer churn, imports the obtained customer churn primary judgment results, the complaint situation and transaction satisfaction data of each transaction into the deep learning churn probability judgment strategy to make a customer churn probability judgment, obtains the judged customer churn probability and compares it with the set customer churn probability threshold, thereby making a judgment on customer churn, and making a customer churn early warning based on the customer churn judgment result, which can help enterprises to promptly discover signs of possible customer churn.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing systems for management, and more specifically, to an automatic early warning method and device for customer churn based on AI machine learning. Background Art

[0002] In business activities, customer churn is an important issue that needs attention. Customer churn refers to the part of customers that a company loses within a certain period of time. For a company, customer churn not only means a reduction in revenue, but also may mean a loss of market share and a decline in brand reputation. Therefore, it is crucial for companies to detect customer churn. Existing technologies cannot conduct a comprehensive analysis of customer sales record data before customer churn, customer complaints for each transaction, and transaction satisfaction data, and the misjudgment rate is high:

[0003] For example, a Chinese patent with publication number CN114662758A discloses a corporate customer loss warning method and device, wherein the method includes: collecting all information about corporate customers; extracting entities, relationships between entities, and entity attributes from the full information about corporate customers; integrating the extracted entities, relationships between entities, and entity attributes to establish a knowledge graph of corporate customers; extracting capital flow information about corporate customers based on the knowledge graph of corporate customers; predicting whether a customer is losing based on the capital flow information about corporate customers, and when the prediction result is that a corporate customer is losing, issuing an alarm message and performing a retention operation. This method can display the capital flow of corporate customers in multiple dimensions and levels through a knowledge graph, and can timely predict customer loss by analyzing the capital flow of corporate customers, thereby improving warning efficiency;

[0004] At the same time, for example, a Chinese patent with publication number CN116091185A provides a customer churn warning method and device, which involves artificial intelligence. The method includes: obtaining customer portrait information and customer behavior information; establishing a BP neural network model, and presetting an expected output sample set; training the BP neural network model based on the customer portrait information and customer behavior information to obtain a customer churn warning model and determine the actual output result; determining customer churn risk information based on the preset expected output sample set and the actual output result. The invention can predict the potential risk of customer churn in the business, so that the account manager can take countermeasures to reduce customer churn in advance and reduce the probability of churn.

[0005] The above patents all have the problems raised by this background technology: the existing technology is unable to conduct a comprehensive analysis of customer sales record data before customer churn, customer complaints for each transaction, and transaction satisfaction data, and the misjudgment rate is high. In order to solve these problems, this application designs an automatic customer churn warning method and device based on AI machine learning. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes an automatic customer churn early warning method and device based on AI machine learning. The present invention sets a judgment cycle, obtains the sales record data of the customer in the enterprise, and simultaneously obtains the customer's complaint situation and transaction satisfaction data for each transaction, obtains the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and imports them into the customer churn primary judgment strategy to make a primary judgment on customer churn, imports the obtained customer churn primary judgment results, the complaint situation and transaction satisfaction data of each transaction into the deep learning churn probability judgment strategy to make a customer churn probability judgment, obtains the judged customer churn probability and compares it with the set customer churn probability threshold, thereby making a judgment on customer churn, and making a customer churn early warning based on the customer churn judgment result, which can help enterprises to promptly discover signs of possible customer churn.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The automatic early warning method for customer churn based on AI machine learning includes the following specific steps:

[0009] S1. Set a judgment cycle to obtain the customer's sales record data in the enterprise, and at the same time obtain the customer's complaints and transaction satisfaction data for each transaction;

[0010] S2, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle and importing them into the customer churn primary judgment strategy to perform primary judgment on customer churn;

[0011] S3, importing the obtained customer churn primary judgment results, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to judge the customer churn probability;

[0012] S4, comparing the determined customer churn probability with a set customer churn probability threshold, thereby determining customer churn;

[0013] S5. Provide customer churn warning based on the customer churn judgment results.

[0014] Specifically, S1 includes the following specific steps:

[0015] S11, taking the average period of the customer's historical product purchases as the judgment period, obtaining the customer's sales record data in the company's historical judgment period, wherein the sales record data includes sales commodity category data and sales amount data, and is stored in the first storage module;

[0016] S12, obtaining the number of customer complaints and transaction satisfaction data after each judgment cycle transaction, and storing them in the second storage module;

[0017] S13. Obtain sales volume change curve data of commodities sold in the market, and store the data in a third storage module.

[0018] Specifically, the customer churn primary judgment strategy in S2 includes the following specific steps:

[0019] S21, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and obtaining the sales volume change curve data of the corresponding sales products in the market;

[0020] S22, substituting the acquired sales record data of the historical judgment period and the sales record data of the current judgment period into the first abnormal value calculation formula to calculate the first abnormal value, the first abnormal value calculation formula is: , where n is the number of judgment cycles, is the sales record data of the i-th judgment cycle;

[0021] S23, obtaining the sales volume change curve data of the corresponding sales commodity in the market, substituting the sales volume change curve data of the corresponding sales commodity in the market into the total amount abnormal value calculation formula to calculate the total amount abnormal value, and the total amount abnormal value calculation formula is: ,in, is the sales volume of the corresponding product in the market in the i-th judgment period;

[0022] S24, substituting the calculated first abnormal value and the total abnormal value into the customer churn primary judgment value calculation formula to calculate the customer churn primary judgment value, wherein the customer churn primary judgment value calculation formula is: .

[0023] Specifically, the deep learning churn probability judgment strategy in S3 includes the following specific steps:

[0024] S31, obtaining the calculated primary judgment value of customer churn, and obtaining the complaint situation and transaction satisfaction data of each transaction;

[0025] S32. Substituting the calculated complaint situation and transaction satisfaction data for each transaction into the customer satisfaction value calculation formula to calculate the customer satisfaction value. The customer satisfaction value calculation formula is: , where m is the number of historical transactions, is the number of complaints for the jth transaction, is the customer satisfaction of the jth transaction, where the transaction satisfaction value ranges from 0 to 1;

[0026] S33, obtaining historical customer satisfaction values ​​and historical customer churn primary judgment values, and at the same time obtaining customer churn situations corresponding to historical customers, and constructing a deep learning neural network model based on the historical customer satisfaction values, the historical customer churn primary judgment values, and the customer churn situations corresponding to historical customers, with the inputs being the historical customer satisfaction values ​​and the historical customer churn primary judgment values, and the output being the customer churn situations corresponding to historical customers;

[0027] S34, dividing the extracted historical customer satisfaction values, historical customer churn primary judgment values, and customer churn situation data corresponding to historical customers into 70% parameter training sets and 30% parameter test sets; inputting 70% parameter training sets into the convolutional neural network model for training to obtain an initial deep learning neural network model; using the 30% parameter test set to test the initial deep learning neural network model, outputting an initial convolutional neural network model that meets the highest customer churn situation judgment accuracy as the deep learning neural network model, and the output strategy formula of the specific neurons in the deep learning neural network model is: ,in is the output of the s-term neuron in the t+1 layer, is the connection weight between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, represents the input of the cth neuron in the tth layer, represents the bias of the linear relationship between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, Represents the Sigmoid activation function.

[0028] Specifically, S4 includes the following specific steps:

[0029] S41, obtaining the calculated customer churn probability and comparing it with a set customer churn probability threshold;

[0030] S42. If the calculated customer churn probability is greater than or equal to the set customer churn probability threshold, the customer churn is judged to be abnormal and there is a risk of churn. If the calculated customer churn probability is less than the set customer churn probability threshold, the customer churn is judged to be normal and there is no risk of churn.

[0031] It should be noted here that the customer churn probability threshold here is preferably 60%;

[0032] Specifically, an automatic customer churn warning device based on AI machine learning is implemented based on the above-mentioned automatic customer churn warning method based on AI machine learning, and specifically includes: a data acquisition module, a customer churn primary judgment module, a customer churn probability calculation module, a customer churn judgment module, a customer churn warning module and a control module. The data acquisition module is used to set a judgment period, obtain the customer's sales record data in the enterprise, and at the same time obtain the customer's complaint situation and transaction satisfaction data for each transaction. The customer churn primary judgment module is used to obtain the sales record data of the historical judgment period and the sales record data of the current judgment period, and import them into the customer churn primary judgment strategy to perform a primary judgment on customer churn.

[0033] Specifically, the customer churn probability calculation module is used to import the obtained customer churn primary judgment results, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to perform customer churn probability judgment. The customer churn judgment module is used to obtain the judged customer churn probability and compare it with the set customer churn probability threshold to make a customer churn judgment. The customer churn warning module is used to make a customer churn warning based on the customer churn judgment results.

[0034] Specifically, the control module is used to control the operation of the data collection module, the customer churn primary judgment module, the customer churn probability calculation module, the customer churn judgment module and the customer churn early warning module.

[0035] Specifically, an electronic device includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0036] The processor executes the above-mentioned automatic customer churn warning method based on AI machine learning by calling the computer program stored in the memory.

[0037] Specifically, a computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned automatic customer churn warning method based on AI machine learning.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention sets a judgment cycle, obtains the sales record data of the customer in the enterprise, and simultaneously obtains the customer's complaint situation and transaction satisfaction data for each transaction, obtains the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and imports them into the customer churn primary judgment strategy to perform a primary judgment on customer churn, and imports the obtained customer churn primary judgment results, the complaint situation of each transaction, and the transaction satisfaction data into the deep learning churn probability judgment strategy to perform a customer churn probability judgment, and compares the obtained customer churn probability with the set customer churn probability threshold, thereby performing a judgment on customer churn, and performing a customer churn warning based on the customer churn judgment result, which can help enterprises to promptly discover signs of possible customer churn. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of the automatic early warning method for customer churn based on AI machine learning of the present invention;

[0041] Figure 2 This is a specific flow chart of step S2 of the customer churn automatic early warning method based on AI machine learning of the present invention;

[0042] Figure 3 This is a specific flow chart of step S3 of the customer churn automatic early warning method based on AI machine learning of the present invention;

[0043] Figure 4 This is a schematic diagram of the architecture of the automatic early warning device for customer churn based on AI machine learning in the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0045] Example 1

[0046] See also Figure 1-Figure 3 , an embodiment provided by the present invention: an automatic early warning method for customer churn based on AI machine learning, which comprises the following specific steps:

[0047] S1. Set a judgment cycle to obtain the customer's sales record data in the enterprise, and at the same time obtain the customer's complaints and transaction satisfaction data for each transaction;

[0048] S2, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle and importing them into the customer churn primary judgment strategy to perform primary judgment on customer churn;

[0049] S3, importing the obtained customer churn primary judgment results, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to judge the customer churn probability;

[0050] S4, comparing the determined customer churn probability with a set customer churn probability threshold, thereby determining customer churn;

[0051] S5. Provide customer churn warning based on the customer churn judgment result;

[0052] In this embodiment, S1 includes the following specific steps:

[0053] S11, taking the average period of the customer's historical product purchases as the judgment period, obtaining the customer's sales record data in the company's historical judgment period, wherein the sales record data includes sales commodity category data and sales amount data, and is stored in the first storage module;

[0054] S12, obtaining the number of customer complaints and transaction satisfaction data after each judgment cycle transaction, and storing them in the second storage module;

[0055] S13, obtaining sales volume change curve data of commodities sold in the market, and storing the data in a third storage module;

[0056] In this embodiment, the primary customer churn determination strategy in S2 includes the following specific steps:

[0057] S21, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and obtaining the sales volume change curve data of the corresponding sales products in the market;

[0058] S22, substituting the acquired sales record data of the historical judgment period and the sales record data of the current judgment period into the first abnormal value calculation formula to calculate the first abnormal value, the first abnormal value calculation formula is: , where n is the number of judgment cycles, is the sales record data of the i-th judgment cycle;

[0059] S23, obtaining the sales volume change curve data of the corresponding sales commodity in the market, substituting the sales volume change curve data of the corresponding sales commodity in the market into the total amount abnormal value calculation formula to calculate the total amount abnormal value, and the total amount abnormal value calculation formula is: ,in, is the sales volume of the corresponding product in the market in the i-th judgment period;

[0060] S24, substituting the calculated first abnormal value and the total abnormal value into the customer churn primary judgment value calculation formula to calculate the customer churn primary judgment value, wherein the customer churn primary judgment value calculation formula is: ;

[0061] In this embodiment, the deep learning churn probability judgment strategy in S3 includes the following specific steps:

[0062] S31, obtaining the calculated primary judgment value of customer churn, and obtaining the complaint situation and transaction satisfaction data of each transaction;

[0063] S32. Substituting the calculated complaint situation and transaction satisfaction data for each transaction into the customer satisfaction value calculation formula to calculate the customer satisfaction value. The customer satisfaction value calculation formula is: , where m is the number of historical transactions, is the number of complaints for the jth transaction, is the customer satisfaction of the jth transaction, where the transaction satisfaction value ranges from 0 to 1;

[0064] S33, obtaining historical customer satisfaction values ​​and historical customer churn primary judgment values, and at the same time obtaining customer churn situations corresponding to historical customers, and constructing a deep learning neural network model based on the historical customer satisfaction values, the historical customer churn primary judgment values, and the customer churn situations corresponding to historical customers, with the inputs being the historical customer satisfaction values ​​and the historical customer churn primary judgment values, and the output being the customer churn situations corresponding to historical customers;

[0065] S34, dividing the extracted historical customer satisfaction values, historical customer churn primary judgment values, and customer churn situation data corresponding to historical customers into 70% parameter training sets and 30% parameter test sets; inputting 70% parameter training sets into the convolutional neural network model for training to obtain an initial deep learning neural network model; using the 30% parameter test set to test the initial deep learning neural network model, outputting an initial convolutional neural network model that meets the highest customer churn situation judgment accuracy as the deep learning neural network model, and the output strategy formula of the specific neurons in the deep learning neural network model is: ,in is the output of the s-term neuron in the t+1 layer, is the connection weight between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, represents the input of the cth neuron in the tth layer, represents the bias of the linear relationship between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, Represents the Sigmoid activation function;

[0066] In this embodiment, S4 includes the following specific steps:

[0067] S41, obtaining the calculated customer churn probability and comparing it with a set customer churn probability threshold;

[0068] S42. If the calculated customer churn probability is greater than or equal to the set customer churn probability threshold, the customer churn is judged to be abnormal and there is a risk of churn. If the calculated customer churn probability is less than the set customer churn probability threshold, the customer churn is judged to be normal and there is no risk of churn.

[0069] It should be noted here that the customer churn probability threshold here is preferably 60%;

[0070] This embodiment can achieve the following: setting a judgment cycle, obtaining the sales record data of customers in the enterprise, and obtaining the customer's complaint situation and transaction satisfaction data for each transaction, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and importing them into the customer churn primary judgment strategy to make a primary judgment on customer churn, importing the obtained customer churn primary judgment results, the complaint situation and transaction satisfaction data of each transaction into the deep learning churn probability judgment strategy to make a customer churn probability judgment, obtaining the judged customer churn probability and comparing it with the set customer churn probability threshold, so as to make a judgment on customer churn, and making a customer churn warning based on the customer churn judgment result, which can help enterprises to promptly discover signs of possible customer churn.

[0071] Example 2

[0072] like Figure 4As shown, the automatic warning device for customer churn based on AI machine learning is implemented based on the above-mentioned automatic warning method for customer churn based on AI machine learning, and specifically includes: a data acquisition module, a customer churn primary judgment module, a customer churn probability calculation module, a customer churn judgment module, a customer churn warning module and a control module, the data acquisition module is used to set the judgment cycle, obtain the sales record data of the customer in the enterprise, and at the same time obtain the customer's complaint situation and transaction satisfaction data for each transaction, the customer churn primary judgment module is used to obtain the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and import them into the customer churn primary judgment strategy to make a primary judgment on customer churn; the customer churn probability calculation module is used to import the obtained customer churn primary judgment result, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to make a customer churn probability judgment, the customer churn judgment module is used to obtain the judged customer churn probability and compare it with the set customer churn probability threshold, so as to make a judgment on customer churn, and the customer churn warning module is used to make a customer churn warning according to the judgment result of customer churn; the control module is used to control the operation of the data acquisition module, the customer churn primary judgment module, the customer churn probability calculation module, the customer churn judgment module and the customer churn warning module.

[0073] Example 3

[0074] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0075] The processor executes the above-mentioned automatic customer churn warning method based on AI machine learning by calling the computer program stored in the memory.

[0076] The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, wherein the memory stores at least one computer program, and the computer program is loaded and executed by the processor to implement the automatic customer churn warning method based on AI machine learning provided by the above method embodiment. The electronic device may also include other components for realizing the functions of the device, for example, the electronic device may also have components such as a wired or wireless network interface and an input and output interface to input and output data. This embodiment will not be described in detail here.

[0077] Example 4

[0078] This embodiment provides a computer-readable storage medium having a rewritable computer program stored thereon;

[0079] When the computer program runs on a computer device, the computer device executes the above-mentioned automatic customer churn warning method based on AI machine learning.

[0080] For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0081] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0082] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0083] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center through a wired network or / and a wireless network. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0084] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0085] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

Claims

1. An automatic early warning method for customer churn based on AI machine learning, characterized in that: It includes the following specific steps: S1. Set a judgment cycle to obtain the customer's sales record data in the enterprise, and at the same time obtain the customer's complaints and transaction satisfaction data for each transaction; S2, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle and importing them into the customer churn primary judgment strategy to perform primary judgment on customer churn; S3, importing the obtained customer churn primary judgment results, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to judge the customer churn probability; S4, comparing the determined customer churn probability with a set customer churn probability threshold, thereby determining customer churn; S5. Provide customer churn warning based on the customer churn judgment result; The customer churn primary judgment strategy in S2 includes the following specific steps: S21, obtaining the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and obtaining the sales volume change curve data of the corresponding sales products in the market; S22, substituting the acquired sales record data of the historical judgment period and the sales record data of the current judgment period into the first abnormal value calculation formula to calculate the first abnormal value, the first abnormal value calculation formula is: , where n is the number of judgment cycles, is the sales record data of the i-th judgment cycle; S23, obtaining the sales volume change curve data of the corresponding sales commodity in the market, substituting the sales volume change curve data of the corresponding sales commodity in the market into the total amount abnormal value calculation formula to calculate the total amount abnormal value, and the total amount abnormal value calculation formula is: ,in, is the sales volume of the corresponding product in the market in the i-th judgment period; S24, substituting the calculated first abnormal value and the total abnormal value into the customer churn primary judgment value calculation formula to calculate the customer churn primary judgment value, wherein the customer churn primary judgment value calculation formula is: ; The deep learning loss probability judgment strategy in S3 includes the following specific steps: S31, obtaining the calculated primary judgment value of customer churn, and obtaining the complaint situation and transaction satisfaction data of each transaction; S32. Substituting the calculated complaint situation and transaction satisfaction data for each transaction into the customer satisfaction value calculation formula to calculate the customer satisfaction value. The customer satisfaction value calculation formula is: , where m is the number of historical transactions, is the number of complaints for the jth transaction, is the customer satisfaction of the jth transaction; The deep learning loss probability judgment strategy also includes the following specific steps: S33, obtaining historical customer satisfaction values ​​and historical customer churn primary judgment values, and at the same time obtaining customer churn situations corresponding to historical customers, and constructing a deep learning neural network model based on the historical customer satisfaction values, the historical customer churn primary judgment values, and the customer churn situations corresponding to historical customers, with the inputs being the historical customer satisfaction values ​​and the historical customer churn primary judgment values, and the output being the customer churn situations corresponding to historical customers; S34, dividing the extracted historical customer satisfaction values, historical customer churn primary judgment values, and customer churn data corresponding to historical customers into 70% parameter training sets and 30% parameter test sets; inputting 70% parameter training sets into the convolutional neural network model for training to obtain an initial deep learning neural network model; using the 30% parameter test set to test the initial deep learning neural network model, outputting an initial convolutional neural network model that meets the highest customer churn judgment accuracy as the deep learning neural network model, and the output strategy formula of the specific neurons in the deep learning neural network model is: ,in is the output of the s-term neuron in the t+1 layer, is the connection weight between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, represents the input of the cth neuron in the tth layer, represents the bias of the linear relationship between the c-term neuron in the t-th layer and the s-term neuron in the t+1-th layer, Represents the Sigmoid activation function.

2. The automatic early warning method for customer churn based on AI machine learning as claimed in claim 1, characterized in that: The S1 comprises the following specific steps: S11, taking the average period of the customer's historical product purchases as the judgment period, obtaining the customer's sales record data in the company's historical judgment period, wherein the sales record data includes sales commodity category data and sales amount data, and is stored in the first storage module; S12, obtaining the number of customer complaints and transaction satisfaction data after each judgment cycle transaction, and storing them in the second storage module; S13. Obtain sales volume change curve data of commodities sold in the market, and store the data in a third storage module.

3. The automatic early warning method for customer churn based on AI machine learning as claimed in claim 2, characterized in that: The S4 includes the following specific steps: S41, obtaining the calculated customer churn probability and comparing it with a set customer churn probability threshold; S42. If the calculated customer churn probability is greater than or equal to the set customer churn probability threshold, the customer churn is judged to be abnormal and there is a risk of churn. If the calculated customer churn probability is less than the set customer churn probability threshold, the customer churn is judged to be normal and there is no risk of churn.

4. An automatic early warning device for customer churn based on AI machine learning, which is implemented based on the automatic early warning method for customer churn based on AI machine learning as claimed in any one of claims 1 to 3, characterized in that: It specifically includes: a data collection module, a customer churn primary judgment module, a customer churn probability calculation module, a customer churn judgment module, a customer churn warning module and a control module. The data collection module is used to set the judgment cycle, obtain the customer's sales record data in the enterprise, and at the same time obtain the customer's complaint situation and transaction satisfaction data for each transaction. The customer churn primary judgment module is used to obtain the sales record data of the historical judgment cycle and the sales record data of the current judgment cycle, and import them into the customer churn primary judgment strategy to perform a primary judgment on customer churn.

5. The automatic early warning device for customer churn based on AI machine learning as claimed in claim 4, characterized in that: The customer churn probability calculation module is used to import the obtained customer churn primary judgment results, the complaint situation of each transaction and the transaction satisfaction data into the deep learning churn probability judgment strategy to perform customer churn probability judgment. The customer churn judgment module is used to obtain the judged customer churn probability and compare it with the set customer churn probability threshold to judge customer churn. The customer churn warning module is used to issue a customer churn warning based on the customer churn judgment results.

6. The automatic early warning device for customer churn based on AI machine learning as claimed in claim 5, characterized in that: The control module is used to control the operation of the data acquisition module, the customer churn primary judgment module, the customer churn probability calculation module, the customer churn judgment module and the customer churn early warning module.

7. An electronic device comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; It is characterized in that the processor executes the automatic customer churn warning method based on AI machine learning as described in any one of claims 1 to 3 by calling the computer program stored in the memory.

8. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes the automatic customer churn warning method based on AI machine learning as described in any one of claims 1 to 3.

Images