Communication user off-network early warning intervention method and system, storage medium and terminal

Abstract

The invention provides a communication user off-network early warning intervention method and system, a storage medium and a terminal. The method comprises the following steps: obtaining user characteristics of a communication user based on multi-source heterogeneous data of the communication user; for the user characteristics, obtaining an off-network prediction probability of the communication user based on a communication user off-network prediction model, and obtaining a high-risk off-network user based on the off-network prediction probability; for the high-risk off-network user, acquiring an off-network cause probability of the high-risk off-network user based on a communication user off-network cause analysis model; and for the user characteristics, the historical reaching feedback data of the high-risk off-network user and the off-network cause probability, obtaining an optimal reaching channel based on a communication user reaching channel recommendation model. According to the communication user off-network early warning and intervention method and system, the storage medium and the terminal, communication user off-network early warning and intervention are carried out based on multi-stage model combination, and the retention success rate and the user satisfaction degree are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of deep learning technology, and in particular to a method, system, storage medium, and terminal for early warning and intervention of communication users leaving the network. Background Technology

[0002] Predicting customer churn is a core task for telecom operators in customer relationship management and reducing customer loss. Currently, mainstream prediction technologies are primarily based on machine learning models, and their evolution path is as follows:

[0003] (1) Early traditional model

[0004] Models such as logistic regression and decision trees primarily rely on basic user attributes (e.g., age, plan) and single consumption statistics (e.g., average revenue per user (ARPU) and call duration). These models are simple but have limited predictive accuracy.

[0005] (2) Current mainstream models

[0006] Ensemble learning algorithms such as XGBoost and LightGBM, while improving prediction accuracy by incorporating more dimensions of user behavior data (such as traffic usage and business subscription records) through feature engineering, are essentially still single-stage prediction models based on a single, static feature set.

[0007] (3) Frontier Exploration Model

[0008] Recent studies have attempted to introduce graph neural networks to analyze users' social relationships or use deep neural networks to process more complex features, but most of these methods still aim to build a more powerful "monotype model" and have not fundamentally changed the prediction paradigm.

[0009] In summary, existing technologies (whether traditional or cutting-edge models) generally share a common fundamental flaw: a single predictive dimension, specifically manifested as follows:

[0010] a. Single feature dimension

[0011] Existing predictive models heavily rely on easily processed structured data (such as tabular data with clearly defined fields like user plans, spending amounts, call durations, and data usage). However, telecom operators possess a wealth of unstructured text data that reflects user sentiment and churn intentions earlier and more directly, yet this data remains underutilized, such as customer service complaints and marketing outreach feedback.

[0012] b. Single output result

[0013] Prediction results are usually just a churn probability value or a binary label. While this can identify "high-risk users," it cannot reveal the specific churn stage a user is currently in or the main driving factors. This prevents operators from developing stage-specific, precise intervention and retention strategies, resulting in costly and inefficient retention efforts. Summary of the Invention

[0014] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a communication user churn early warning and intervention method, system, storage medium and terminal, which jointly performs early warning and intervention for communication user churn based on a multi-stage model, significantly improving the retention success rate and user satisfaction.

[0015] In a first aspect, the present invention provides a method for early warning and intervention of communication user churn, the method comprising the following steps: obtaining user characteristics of the communication user based on multi-source heterogeneous data of the communication user; obtaining the churn prediction probability of the communication user based on the user characteristics and the churn prediction probability to identify high-risk churn users; obtaining the churn motivation probability of the high-risk churn users based on the churn motivation analysis model; and obtaining the optimal reach channel based on the user characteristics, the historical reach feedback data of the high-risk churn users, and the churn motivation probability, using a communication user reach channel recommendation model.

[0016] In one implementation of the first aspect, obtaining the user characteristics of the communication user based on multi-source heterogeneous data of the communication user includes the following steps:

[0017] Obtain multi-source heterogeneous data of the communication user, including the communication user's package usage, consumption bills, call details, customer service complaint texts, and communication quality data;

[0018] Extract statistical features based on the usage of the package, the consumption bill, and the call details as time-series features;

[0019] Textual features of the customer complaint text were extracted based on the BERT model;

[0020] Based on the communication quality data, communication interaction features are extracted;

[0021] The user features are constructed based on the time-series features, the text features, and the communication interaction features.

[0022] In one implementation of the first aspect, the steps of obtaining the predicted churn probability of the communication user based on the communication user churn prediction model, and obtaining high-risk churn users based on the predicted churn probability, include the following steps:

[0023] The user characteristics are input into the communication user churn prediction model to obtain the churn prediction probability of the communication user within a preset period.

[0024] Determine whether the predicted churn probability is greater than a preset threshold; if so, determine that the communication user is a candidate high-risk churn user.

[0025] Sort all candidate high-risk churn users by their predicted churn probability from largest to smallest, and select the top preset number of candidate high-risk churn users as the high-risk churn users.

[0026] In one implementation of the first aspect, obtaining the probability of churn motivation for the high-risk churned user based on the communication user churn motivation analysis model includes the following steps:

[0027] Obtain the churn warning characteristics of the high-risk churned users, including poor network quality characteristics, changes in SIM card slots on other networks, and marketing contact characteristics from other networks;

[0028] The churn warning features are input into the communication user churn motivation analysis model to obtain the probability of each churn motivation, wherein the churn motivation includes tariff, network quality, service quality and competitor attraction.

[0029] In one implementation of the first aspect, obtaining the optimal reach channel based on a communication user reach channel recommendation model, taking into account the user characteristics, the historical reach feedback data of the high-risk churned users, and the probability of churn motivation, includes the following steps:

[0030] Based on the historical contact feedback data of the high-risk disconnected users, the corresponding contact channels are obtained;

[0031] The user characteristics, the reach channels, and the probability of churn are input into the communication user reach channel recommendation model to obtain the conversion probability of different reach channels;

[0032] The channel with the highest conversion probability is selected as the optimal channel.

[0033] In one implementation of the first aspect, the reach channels include telephone, SMS, and APP push.

[0034] In one implementation of the first aspect, the communication user churn prediction model adopts an XGBoost model, a LightGBM model, or a Stacking ensemble model; the communication user churn motivation analysis model adopts a multi-task learning model or a multi-classification model; and the communication user reach channel recommendation model adopts a gradient boosting tree or a neural network based on few-shot learning.

[0035] Secondly, the present invention provides a communication user churn early warning and intervention system, the system comprising an acquisition module, a prediction module, an analysis module and a recommendation module;

[0036] The acquisition module is used to acquire the user characteristics of the communication user based on the multi-source heterogeneous data of the communication user;

[0037] The prediction module is used to obtain the predicted churn probability of the communication user based on the communication user churn prediction model based on the user characteristics, and to obtain high-risk churn users based on the predicted churn probability.

[0038] The analysis module is used to obtain the probability of churn motivation for the high-risk churn users based on the communication user churn motivation analysis model.

[0039] The recommendation module is used to obtain the optimal reach channel based on the user characteristics, the historical reach feedback data of the high-risk churned users, and the probability of churn motivation, using a communication user reach channel recommendation model.

[0040] Thirdly, the present invention provides a storage medium on which a computer program is stored, which, when executed by a processor, implements the above-mentioned communication user disconnection early warning intervention method.

[0041] Fourthly, the present invention provides a terminal, comprising: a processor and a memory;

[0042] The memory is used to store computer programs;

[0043] The processor is used to execute the computer program stored in the memory, so that the terminal performs the above-described communication user churn warning intervention method.

[0044] As described above, the communication user churn early warning intervention method, system, medium, and device of the present invention have the following beneficial effects:

[0045] (1) The discrete communication user churn prediction, analysis and decision-making processes are integrated into an automated intelligent pipeline, which greatly improves operational efficiency.

[0046] (2) It not only answers "who will leave the network", but also answers "why they leave the network" and "how to retain them", making the retention strategy change from "one-size-fits-all" to "personalized", significantly improving the retention success rate and user satisfaction.

[0047] (3) Through the joint design and information transmission of the three-stage model, deep reasoning capabilities are formed. For example, the churn motivation is input as a key feature into the outreach channel recommendation model, making the selection of outreach channels more targeted (such as recommending discounts to users who are sensitive to tariffs, and prioritizing communication with account managers for users who are dissatisfied with the service).

[0048] (4) By recommending channels accurately, we avoided blindly reaching out, and improved the conversion rate while controlling retention costs. Attached Figure Description

[0049] Figure 1 The flowchart shown is an embodiment of the communication user churn early warning intervention method of the present invention;

[0050] Figure 2 The diagram shows a framework schematic of one embodiment of the communication user churn early warning intervention method of the present invention;

[0051] Figure 3 The diagram shown is a structural schematic of the communication user churn early warning and intervention system of the present invention in one embodiment;

[0052] Figure 4 The diagram shown is a structural schematic of the terminal of the present invention in one embodiment. Detailed Implementation

[0053] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0054] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0055] The following embodiments of the present invention provide a method for early warning and intervention of communication users leaving the network, which can be applied to terminals. The terminals mentioned in the present invention may include mobile phones, tablet computers, laptops, wearable devices, in-vehicle devices, augmented reality (AR) / virtual reality (VR) devices, ultra-mobile personal computers (UMPCs), netbooks, personal digital assistants (PDAs), etc., with no restrictions on the specific type of terminal in the embodiments of the present invention.

[0056] For example, the terminal may be a station (ST) in a WLAN with wireless charging capability, a cellular phone, cordless phone, Session Initiation Protocol (SIP) phone, Wireless Local Loop (WLL) station, Personal Digital Assistant (PDA) device, handheld device with wireless charging capability, computing device or other processing device, computer, laptop computer, handheld communication device, handheld computing device, and / or other devices for communication over a wireless system, as well as next-generation communication systems, such as mobile terminals in 5G networks, mobile terminals in future evolved Public Land Mobile Networks (PLMNs), or mobile terminals in future evolved Non-terrestrial Networks (NTNs).

[0057] For example, the terminal can communicate with networks and other devices wirelessly. The wireless communication can use any communication standard or protocol, including but not limited to Global System for Mobile Communication (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), email, Short Messaging Service (SMS), BT, GNSS, WLAN, NFC, FM, and / or IR technologies. The GNSS can include Global Positioning System (GPS), Global Navigation Satellite System (GLONASS), BeiDou Navigation Satellite System (BDS), Quasi-Zenith Satellite System (QZSS), and / or Satellite Based Augmentation Systems (SBAS).

[0058] The technical solutions of the present invention will now be described in detail with reference to the accompanying drawings.

[0059] like Figure 1 and Figure 2 As shown, in one embodiment, the communication user churn warning intervention method of the present invention includes steps S1-S4.

[0060] Step S1: Obtain the user characteristics of the communication user based on the multi-source heterogeneous data of the communication user.

[0061] Specifically, obtaining the user characteristics of the communication user based on multi-source heterogeneous data of the communication user includes the following steps:

[0062] 11) Obtain multi-source heterogeneous data of the communication user, including the communication user's package usage, consumption bills, call details, customer service complaint texts, and communication quality data.

[0063] In order to protect the privacy of the communication users, the multi-source heterogeneous data can be anonymized first.

[0064] 12) Extract statistical features based on the usage of the package, the consumption bill, and the call details as time-series features.

[0065] For example, the mean, variance, and recent trends (such as the month-on-month decrease rate of call volume in the past 7 days) of the communication user's consumption amount, call duration, and data usage over the past 1, 3, and 6 months are calculated as the time series features.

[0066] 13) Extract text features from the customer complaint text based on the BERT model.

[0067] Specifically, complaint sentiment features and complaint demand features are extracted from the customer complaint text as the text features.

[0068] 14) Extract communication interaction features based on the communication quality data.

[0069] For example, by combining user location information with communication quality data, communication interaction characteristics such as "poor network quality in the resident area" can be generated.

[0070] 15) Construct the user features based on the time sequence features, the text features, and the communication interaction features.

[0071] Step S2: Based on the user characteristics, obtain the churn prediction probability of the communication user based on the communication user churn prediction model, and obtain high-risk churn users based on the churn prediction probability.

[0072] Specifically, a communication user churn prediction model is pre-trained to predict the probability of communication users churning. The user characteristics are input into the communication user churn prediction model to obtain the predicted churn probability of the communication user within a preset period (e.g., one month). It is determined whether the predicted churn probability is greater than a preset threshold (e.g., 0.7). If so, the communication user is identified as a candidate high-risk churn user. For the filtered candidate high-risk churn users, their predicted churn probabilities are sorted from largest to smallest, and the top K candidate high-risk churn users are selected as the high-risk churn users. The preset number K can be determined based on the retention resource budget for the current operating period, such as the number of customer service seats or the upper limit of the marketing budget.

[0073] Preferably, the communication user churn prediction model adopts the XGBoost (Extreme Gradient Boosting) model, the LightGBM (Light Gradient Boosting Machine) model, or the Stacking ensemble model.

[0074] Step S3: For the high-risk churned users, obtain the probability of churn motivation for the high-risk churned users based on the communication user churn motivation analysis model.

[0075] Specifically, in this invention, churn motivations include tariffs, network quality, service quality, and competitor attraction. A communication user motivation analysis module is pre-trained to obtain the probability that a communication user belongs to each churn motivation. For high-risk churn users, churn warning characteristics are first obtained, including poor network quality characteristics, SIM card slot change characteristics, and marketing contact characteristics from other networks. The poor network quality characteristics are used to quantify the degree of user experience degradation when using the operator's network. Essentially, it extracts indicators reflecting network stability, coverage quality, and service smoothness through base station performance data, terminal signaling data, and customer service complaint tickets. The SIM card slot change characteristics are based on changes in the SIM card configuration of the user's mobile terminal SIM card slot, capturing behavioral signals of the user attempting to switch networks. It is a key transition indicator from "willingness to switch networks" to "action to switch networks." Essentially, it determines whether the user has inserted a SIM card from another network and whether they have switched between primary and secondary SIM cards through terminal-side SIM card slot identification data and signaling monitoring data. The aforementioned cross-network marketing contact characteristics are used to characterize the extent to which competitor operators reach and penetrate the network's users. This is a core external factor that amplifies users' willingness to switch networks. Essentially, it extracts indicators of users receiving cross-network marketing, paying attention to cross-network packages, and participating in cross-network activities through user behavior data and channel feedback data. Then, the cross-network churn warning characteristics are input into the communication user churn motivation analysis model to obtain the probability of each churn motivation, such as: [Tariff: 0.6, Network: 0.3, Service: 0.1].

[0076] Preferably, the communication user churn motivation analysis model adopts a multi-task learning model or a multi-classification model.

[0077] Step S4: Based on the user characteristics, the historical contact feedback data of the high-risk churned users, and the probability of churn motivation, obtain the optimal contact channel based on the communication user contact channel recommendation model.

[0078] Specifically, a communication user outreach channel recommendation model is pre-trained to output the conversion probability of communication users for different outreach channels. The outreach channel refers to the carrier and path through which marketing information (such as package discounts, network switching warnings and retention efforts, and new product promotions) is delivered to users, including telephone, SMS, and app push notifications. For high-risk churn users, the corresponding outreach channels are first obtained based on their historical outreach feedback data (such as historical retention records, outreach channels, outreach scripts, retention success indicators, historical marketing channels, number of contacts with the user, and number of successes). Then, the user characteristics, the outreach channels, and the probability of churn motivation are input into the communication user outreach channel recommendation model to obtain the conversion probability of different outreach channels. Finally, the outreach channel with the highest conversion probability is selected as the optimal outreach channel to formulate a personalized retention strategy.

[0079] Preferably, the communication user access channel recommendation model adopts a gradient boosting tree or a neural network based on few-shot learning.

[0080] Therefore, this invention uses a pipeline-style collaborative operation of multi-stage models, with each subsequent model depending on the output of the previous model, to collectively form a complete communication user churn early warning and intervention system.

[0081] The protection scope of the communication user churn early warning intervention method described in this embodiment is not limited to the execution order of the steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principle of this invention is included within the protection scope of this invention.

[0082] This invention also provides a communication user churn warning and intervention system. The communication user churn warning and intervention system can implement the communication user churn warning and intervention method described in this invention. However, the implementation device of the communication user churn warning and intervention system described in this invention includes, but is not limited to, the structure of the communication user churn warning and intervention system listed in this embodiment. All structural modifications and substitutions of the prior art made according to the principles of this invention are included within the protection scope of this invention.

[0083] like Figure 3 As shown, in one embodiment, the communication user churn early warning intervention system of the present invention includes an acquisition module 31, a prediction module 32, an analysis module 33, and a recommendation module 34.

[0084] The acquisition module 31 is used to acquire the user characteristics of the communication user based on the multi-source heterogeneous data of the communication user.

[0085] The prediction module 32 is connected to the acquisition module 31 and is used to obtain the churn prediction probability of the communication user based on the communication user churn prediction model based on the user characteristics, and to obtain high-risk churn users based on the churn prediction probability.

[0086] The analysis module 33 is connected to the prediction module 32 and is used to obtain the probability of the churn motivation of the high-risk churn users based on the communication user churn motivation analysis model.

[0087] The recommendation module 34 is connected to the analysis module 33 and is used to obtain the optimal reach channel based on the communication user reach channel recommendation model, taking into account the user characteristics, the historical reach feedback data of the high-risk churned users and the probability of churn motivation.

[0088] The structure and principle of the acquisition module 31, prediction module 32, analysis module 33 and recommendation module 34 correspond one-to-one with the steps in the above-mentioned communication user churn early warning intervention method, so they will not be described again here.

[0089] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0090] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of the present invention, depending on actual needs. For example, the functional modules / units in the various embodiments of the present invention may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0091] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0092] This invention also provides a computer-readable storage medium. Those skilled in the art will understand that all or part of the steps in the communication user churn warning intervention method of the above embodiments can be executed by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. This available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0093] This invention also provides a terminal. The terminal includes a processor and a memory.

[0094] The memory is used to store computer programs.

[0095] The memory includes various media capable of storing program code, such as ROM, RAM, magnetic disk, USB flash drive, memory card, or optical disk.

[0096] The processor is connected to the memory and is used to execute the computer program stored in the memory, so that the terminal performs the above-mentioned communication user disconnection early warning intervention method.

[0097] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0098] like Figure 4 As shown, the terminal of the present invention is presented in the form of a general-purpose computing device. The components of the terminal may include, but are not limited to: one or more processors or processing units 41, a memory 42, and a bus 43 connecting different system components (including the memory 42 and the processing unit 41).

[0099] Bus 43 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0100] Terminals typically include various computer system-readable media. These media can be any available media that can be accessed by the terminal, including volatile and non-volatile media, and removable and non-removable media.

[0101] Memory 42 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 421 and / or cache memory 422. The terminal may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 423 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 4 Not shown; usually referred to as a "hard drive"). Although Figure 4Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 43 via one or more data media interfaces. Memory 42 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0102] A program / utility 424 having a set (at least one) of program modules 4241 may be stored, for example, in memory 42. Such program modules 4241 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 4241 typically perform the functions and / or methods described in the embodiments of the present invention.

[0103] The terminal can also communicate with one or more external devices (e.g., keyboard, pointing device, display, etc.), one or more devices that enable user interaction with the terminal, and / or any device that enables the terminal to communicate with one or more other computing devices (e.g., network interface card, modem, etc.). This communication can be performed through input / output (I / O) interface 44. Furthermore, the terminal can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) through network adapter 45. Figure 4 As shown, network adapter 45 communicates with other modules of the terminal via bus 43. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the terminal, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0104] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for early warning and intervention of communication user churn, characterized in that, The method includes the following steps: The user characteristics of the communication users are obtained based on the multi-source heterogeneous data of the communication users; Based on the user characteristics, the churn prediction probability of the communication user is obtained based on the communication user churn prediction model, and high-risk churn users are identified based on the churn prediction probability. For the high-risk churned users, the probability of churn motivation for the high-risk churned users is obtained based on the communication user churn motivation analysis model; Based on the user characteristics, the historical contact feedback data of the high-risk churned users, and the probability of churn motivation, the optimal contact channel is obtained based on the communication user contact channel recommendation model.

2. The communication user churn early warning and intervention method according to claim 1, characterized in that, Obtaining the user characteristics of a communication user based on multi-source heterogeneous data of the communication user includes the following steps: Obtain multi-source heterogeneous data of the communication user, including the communication user's package usage, consumption bills, call details, customer service complaint texts, and communication quality data; Extract statistical features based on the usage of the package, the consumption bill, and the call details as time-series features; Textual features of the customer complaint text were extracted based on the BERT model; Based on the communication quality data, communication interaction features are extracted; The user features are constructed based on the time-series features, the text features, and the communication interaction features.

3. The communication user churn early warning and intervention method according to claim 1, characterized in that, Based on the aforementioned user characteristics, the following steps are taken to obtain the predicted churn probability of the communication user according to the communication user churn prediction model, and to identify high-risk churn users based on the predicted churn probability: The user characteristics are input into the communication user churn prediction model to obtain the churn prediction probability of the communication user within a preset period. Determine whether the predicted churn probability is greater than a preset threshold; if so, determine that the communication user is a candidate high-risk churn user. Sort all candidate high-risk churn users by their predicted churn probability from largest to smallest, and select the top preset number of candidate high-risk churn users as the high-risk churn users.

4. The communication user churn early warning intervention method according to claim 1, characterized in that, For the high-risk churned users, the probability of churn motivation for these users, based on the communication user churn motivation analysis model, includes the following steps: Obtain the churn warning characteristics of the high-risk churned users, including poor network quality characteristics, changes in SIM card slots on other networks, and marketing contact characteristics from other networks; The churn warning features are input into the communication user churn motivation analysis model to obtain the probability of each churn motivation, wherein the churn motivation includes tariff, network quality, service quality and competitor attraction.

5. The communication user churn early warning intervention method according to claim 1, characterized in that, Based on the user characteristics, historical contact feedback data of the high-risk churned users, and the probability of churn motivation, obtaining the optimal contact channel based on the communication user contact channel recommendation model includes the following steps: Based on the historical contact feedback data of the high-risk disconnected users, the corresponding contact channels are obtained; The user characteristics, the reach channels, and the probability of churn are input into the communication user reach channel recommendation model to obtain the conversion probability of different reach channels; The channel with the highest conversion probability is selected as the optimal channel.

6. The communication user churn early warning intervention method according to claim 5, characterized in that, The channels for reaching out include telephone, SMS, and app push notifications.

7. The communication user churn early warning intervention method according to claim 1, characterized in that, The communication user churn prediction model adopts the XGBoost model, LightGBM model, or Stacking ensemble model; the communication user churn motivation analysis model adopts the multi-task learning model or multi-classification model; and the communication user reach channel recommendation model adopts the gradient boosting tree or a neural network based on few-shot learning.

8. A communication user churn early warning and intervention system, characterized in that, The system includes an acquisition module, a prediction module, an analysis module, and a recommendation module; The acquisition module is used to acquire the user characteristics of the communication user based on the multi-source heterogeneous data of the communication user; The prediction module is used to obtain the predicted churn probability of the communication user based on the communication user churn prediction model based on the user characteristics, and to identify high-risk churn users based on the predicted churn probability. The analysis module is used to obtain the probability of churn motivation for the high-risk churn users based on the communication user churn motivation analysis model. The recommendation module is used to obtain the optimal reach channel based on the user characteristics, the historical reach feedback data of the high-risk churned users, and the probability of churn motivation, using a communication user reach channel recommendation model.

9. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the communication user disconnection early warning intervention method as described in any one of claims 1 to 7.

10. A terminal, characterized in that, include: Processor and memory; The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory, so that the terminal performs the communication user disconnection early warning intervention method according to any one of claims 1 to 7.

Images