Complaint prediction method

Abstract

Embodiments of the present disclosure provide a complaint prediction method and apparatus. The method comprises: acquiring user identification information, service behavior information, and service index information; on the basis of the user identification information, the service behavior information, and the service index information, determining a service state of a user; and, on the basis of the service state, predicting a user complaint probability.

Description

Complaint prediction method

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This disclosure is based on and claims the priority of Chinese patent application 202311796325.0 filed on December 22, 2023, and all the disclosed contents are incorporated into this disclosure by reference. Technical Field

[0003] The present disclosure relates to the field of communications, and more particularly, to a complaint prediction method and device. Background Art

[0004] In the process of digital transformation across society, facing the digital and intelligent demands of various industries, business types are becoming increasingly diverse and complex. Growing customer demands for differentiated services and the integration of large-scale, complex networks are placing higher demands on operators to provide agile, personalized support for users. As 2C service providers, telecommunications operators must prioritize improving user satisfaction and reducing customer complaints. For existing issues, operators must be able to identify potential complaint risks and prevent them through proactive action.

[0005] Related technologies typically use feature vectors constructed from signaling data, user business domain (B-domain) data, or historical complaint data as data sources and input them into AI models. The AI ​​models then automatically learn user behaviors, feature similarities, or potential connections within them to predict future user complaints. However, the predictive effectiveness of models generated by this approach is often limited by the quality and magnitude of the samples, making it extremely difficult to generate practical value in the communications field's complaint scenarios. Furthermore, the entire prediction process is unclear, and the prediction results are poorly interpretable, making it impossible to guide network optimization personnel in their next steps and generating no practical benefits. Summary of the Invention

[0006] The disclosed embodiments provide a complaint prediction method and device to at least solve the problems in related technologies such as unclear prediction process, poor interpretability of prediction results, inability to guide network optimization personnel to implement the next step of work, and inability to generate practical benefits.

[0007] According to one embodiment of the present disclosure, a complaint prediction method is provided, comprising:

[0008] Obtain user identification information, business behavior information, and business indicator information;

[0009] Determine the user's business status based on user identification information, business behavior information, and business indicator information;

[0010] Predict the probability of user complaints based on business status.

[0011] According to another embodiment of the present disclosure, a prediction device is provided, comprising:

[0012] The acquisition module is used to obtain user identification information, business behavior information, and business indicator information;

[0013] A determination module is used to determine the user's business status based on user identification information, business behavior information, and business indicator information;

[0014] The prediction module is used to predict the probability of user complaints based on business status.

[0015] According to another embodiment of the present disclosure, a computer-readable storage medium is provided, in which a computer program is stored. The computer program is configured to execute the steps of any one of the above method embodiments when running.

[0016] According to another embodiment of the present disclosure, an electronic device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to perform the steps in any one of the above method embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG1 is a hardware structure block diagram of a mobile terminal of a complaint prediction method according to an embodiment of the present disclosure;

[0018] FIG2 is a flowchart of a complaint prediction method according to an embodiment of the present disclosure;

[0019] FIG3 is a schematic diagram of a complaint prediction model according to an embodiment of the present disclosure;

[0020] FIG4 is a structural block diagram of a complaint prediction apparatus according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0021] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings and in conjunction with embodiments.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0023] Domain B data includes user data and business data, such as user consumption habits, terminal information, business content, business target audience, etc.

[0024] Operation support system (O domain) data, including network data, such as signaling, alarms, faults, network resources, etc.

[0025] The existing method uses feature vectors constructed from signaling data, user B-domain data, or historical complaint data as data sources and inputs them into AI models. The AI ​​models then automatically learn user behaviors, feature similarities, or internal potential connections to predict future user complaints. However, the prediction effect is limited by the quality and magnitude of the samples, and the entire prediction process is unclear. The prediction results are poorly interpretable, unable to guide network optimization personnel in implementing the next step of work, and unable to generate practical benefits.

[0026] In response to the above-mentioned technical problems, the present disclosure provides a complaint prediction method. The technical concept is to combine the user's user identification information, business behavior information, and business indicator information, that is, to combine O domain data and B data, to determine the user's business status, and predict the user's complaint probability based on the business status. The algorithm logic is visual and easy to understand. For the predicted complaining users, the reasons for the user's complaint and the subsequent specific solutions can be clearly given, which has better practicality and implementation value. It provides strong support for the operation and maintenance support and operation analysis of mobile operators, while greatly saving labor costs.

[0027] In an embodiment of the present disclosure, B-domain data may include operator B-domain data (e.g., business behavior information), and O-domain data may include user O-domain data information (e.g., user identification information) and O-domain key performance indicator (KPI) information (e.g., business indicator information).

[0028] Exemplarily, the complaint prediction method of an embodiment of the present invention can be implemented by a user complaint prediction model. For example, O-domain data and / or B-domain data can be input into the user complaint prediction model to output a user complaint prediction result.

[0029] The user complaint model can be applied to various sectors providing 2C services to improve service quality, reduce the number of user complaints, and enhance customer satisfaction. These sectors include, but are not limited to, telecommunications or communications, retail and e-commerce, financial services, and the hotel and tourism industry.

[0030] The user complaint prediction model of the embodiment of the present disclosure may be used, but is not limited to, in the following applications:

[0031] (1) Improving operator service quality: In the telecommunications or communications sector, operators can use user complaint prediction models to predict and identify network failures, service interruptions, or performance issues, and take preventive measures to improve service quality and reduce the number of user complaints. For example, network performance can be optimized, and faults can be predicted and resolved.

[0032] (2) Improvement of customer care and satisfaction: Based on the analysis and prediction results of the user complaint prediction model, telecom operators can customize personalized customer care plans to better meet customer needs, improve customer satisfaction, and increase customer loyalty.

[0033] (3) Real-time monitoring of faults and problems: User complaint prediction models can be combined with real-time data to monitor faults and problems in the network and services. This helps operators identify and resolve problems in a timely manner, reducing service interruptions and user dissatisfaction.

[0034] (4) Resource allocation and network optimization: Through user complaint prediction models, operators can allocate resources more efficiently, optimize network performance, and take faster actions in high-complaint risk areas to improve service quality.

[0035] (5) Market competitive advantage: Operators can use user complaint prediction models to identify market trends and user needs in advance, thereby better competing, launching new products and services, and meeting the ever-changing needs of the market.

[0036] (6) Cross-industry application: In addition to the telecommunications and communications fields, the user complaint prediction model can also be applied to other fields, such as retail, customer service, hotel management, etc., to predict customer complaints in advance and improve service quality and customer experience.

[0037] Furthermore, the user complaint prediction model disclosed herein can be combined with other known or unknown technologies to leverage its potential in a wider range of applications. For example, integration with automated machine learning systems, big data analytics, and natural language processing technologies can improve the model's accuracy and efficiency, broadening its application areas. The user complaint prediction model disclosed herein has the potential for widespread application, improving service quality, increasing customer satisfaction, and increasing efficiency and competitiveness in a variety of fields.

[0038] The method embodiments provided in the embodiments of the present disclosure can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 is a hardware structure block diagram of a mobile terminal of a complaint prediction method in an embodiment of the present disclosure. As shown in Figure 1, the mobile terminal may include one or more (only one is shown in Figure 1) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the above-mentioned mobile terminal may also include a transmission device 106 and an input and output device 108 for communication functions. It can be understood by those skilled in the art that the structure shown in Figure 1 is only for illustration and does not limit the structure of the above-mentioned mobile terminal. For example, the mobile terminal may also include more or fewer components than those shown in Figure 1, or have a configuration different from that shown in Figure 1.

[0039] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the complaint prediction method in the embodiment of the present disclosure. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above-mentioned method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0040] The transmission device 106 is used to receive or send data via a network. A specific example of the aforementioned network may include a wireless network provided by the mobile terminal's communications provider. In one embodiment, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0041] In this embodiment, a complaint prediction method is provided. FIG2 is a flow chart of the complaint prediction method according to an embodiment of the present disclosure. As shown in FIG2 , the process includes the following steps:

[0042] Step S201: Acquire user identification information, business behavior information, and business indicator information.

[0043] As an example, user identification information may include information such as the International Mobile Equipment Identity (IMSI), terminal model, cell ID (CI), etc., and user identification information belongs to O-domain data information; business behavior information can be used to evaluate whether the user's business preferences have changed and to judge inconsistent behavior. Business behavior information may include information such as permanent areas, common services, common time periods, etc., and business behavior information belongs to B-domain data information; business indicator information may include business key performance KPIs (such as latency, TCP link establishment success rate, etc.), business key quality indicators KQIs (such as web page opening latency, video download rate, etc.), and cell basic KPIs (such as capacity, coverage, interference, etc.). Business indicator information is network data information and belongs to O-domain data information.

[0044] Step S202: Determine the user's service status based on the user identification information, the service behavior information, and the service indicator information.

[0045] As an example, the service status of a user may include a changed status and an unchanged status, and whether the service status of the user has changed may be determined based on user identification information, service behavior information, and service indicator information.

[0046] As an example, the types of user's business status may include but are not limited to business behavior status and business indicator status, where the business behavior status may include multiple types, such as business preference status, time preference status, resident area preference status, etc.; the business indicator status may include business indicator change status and business indicator unchanged status, where the business indicator change status may include business indicator degradation and business indicator non-degradation.

[0047] As an example, service preferences are a broad concept that can be expanded based on the operator's current network development status. Taking domestic operators as an example, current 5G network coverage can support most business scenarios, so services can be expanded to five categories: web browsing, long video, short video, gaming, instant messaging, and voice.

[0048] In an exemplary embodiment, the business behavior information includes current business behavior information and historical business behavior information; the business status includes business behavior status; the above step S202 includes:

[0049] Step S2021a, determining the business behavior status of the user according to the user identification information and the business behavior information corresponding to the user;

[0050] Step S2022a: When the current business behavior information is inconsistent with the historical business behavior information, determine that the business behavior status is a changed status.

[0051] As an example, the business behavior information may include current business behavior information and historical business behavior information. The current business behavior information and the historical business behavior information of the user may be analyzed. If the current business behavior information and the historical business behavior information of the user are inconsistent, it may be determined that the user's business behavior status is in a changing state.

[0052] As an example, the types of business behavior information may include, but are not limited to, business usage type information, business usage time information, and business usage address information.

[0053] The process of step S2021a and step S2022a is further described below:

[0054] In an exemplary embodiment, the service behavior information includes service usage duration information and service usage type information, and the service behavior status includes a service preference status. Step S2021a may include:

[0055] Analyzing the current service usage duration information and historical service usage duration information of the user for the same type of service based on the user identification information, the service usage duration information, and the service usage type information corresponding to the user, and determining the service preference status of the user;

[0056] Step S2022a may include:

[0057] When the analysis results of the current service usage duration information and the historical service usage duration information of the user for the same type of service are inconsistent, it is determined that the service preference state is a changed state.

[0058] As an example, a change in the terminal user's business behavior can lead to a change in the user's business preferences. Business behavior information can include business usage duration information and business usage type information. The business usage duration and business usage type can be used to identify whether the user's business preferences have changed. For example, if the user's current web browsing time is shorter than the historical web browsing time and the user's current video browsing time is longer than the historical video browsing time, or if the user's historical web browsing time is longer than the historical video browsing time, but the current web browsing time is shorter than the current video browsing time, it can be indicated that the web page preference user has changed to a video preference user.

[0059] As an example, different types of services used by users can be sorted from high to low according to the service usage time at different times, and the current usage time sorting results and historical usage time sorting results of different types of services can be obtained. The user's service preference status can be judged based on the current usage time sorting results and the historical usage time sorting results. When the current usage time sorting results and the historical usage time sorting results are different, it can be determined that the service preference status has changed.

[0060] It should be noted that in addition to determining the service preference status based on the ranking results of the user's service usage time (whether the ranking results have changed), it can also be determined based on other methods, such as determining it based on the proportion of service usage time (whether the proportion of service usage time has changed), or setting a service usage time threshold based on actual conditions, or using an AI algorithm, etc. The embodiments of the present invention do not limit this.

[0061] In addition, other states of the business behavior status, such as the time preference state, the resident area preference state, etc., can also be determined by using one or more of the above determination methods, and the embodiment of the present invention does not limit this.

[0062] As an example, for users whose service preferences have changed, it is possible to identify whether the network indicators corresponding to the new preferred services meet the network usage standards corresponding to the new preferred services, so as to meet the user's usage needs for the changed services.

[0063] As an example, for users whose service preferences have not changed, it is possible to identify whether the current network indicators of the user's preferred service are lower than the corresponding service usage network standards. If the current network indicators of the user's preferred service are lower than the corresponding service usage network standards, it can be considered that the current network indicators of the user's preferred service have deteriorated.

[0064] In an exemplary embodiment, the service behavior information includes service usage period information, and the service behavior status includes a time preference status. Step S2021a may include:

[0065] Analyzing the current service usage time period information of the user and the historical service usage time period information of the user according to the user identification information and the service usage time period information corresponding to the user, and determining the time preference state of the user;

[0066] Step S2022a may include:

[0067] When the analysis results of the current service usage period information of the user and the historical service usage period information are inconsistent, it is determined that the time preference state is a changed state.

[0068] As an example, business behavior information may include business usage period information. After the user's time preference for using the business terminal changes, the network load may be different due to differences in the number of network users, business behavior, etc. in different time periods, which may ultimately cause the user's business perception in the current usage period to differ from the business perception in the historical usage period, thereby causing the user's current perception to be inconsistent with expectations.

[0069] As an example, the user's current business usage period information and historical business usage period information can be analyzed based on the user's corresponding user identification information and business usage period information. If the analysis results of the user's current business usage period information and historical business usage period information are inconsistent, it can be determined that the user's time preference status has changed.

[0070] As an example, for a user whose time preference status has changed, it is possible to further analyze whether the network indicators corresponding to the service usage period after the time preference change have deteriorated.

[0071] In an exemplary embodiment, the service behavior information includes service usage address information, and the service behavior status includes a resident area preference status. Step S2021a may include:

[0072] analyzing the current service usage address information of the user and the historical service usage address information of the user according to the user identification information and the service usage address information corresponding to the user, and determining the resident area preference state of the user;

[0073] Step S2022a may include:

[0074] When the current service usage address information of the user is inconsistent with the historical service usage address information, it is determined that the resident zone preference state of the user is a changed state.

[0075] As an example, service behavior information may include service usage address information, and service behavior status may include resident region preference status. After a user's resident region preference status for a service terminal changes, due to differences in geographical environment, network coverage, network load, etc., the user's current terminal service usage may differ from their historical resident region, resulting in a mismatch between the user's current network usage perception and expectations.

[0076] As an example, the user's current business usage address information and historical business usage address information can be analyzed based on the user's corresponding user identification information and business usage address information. If the user's current business usage address information is inconsistent with the historical business usage address information, it can be determined that the user's resident area preference status has changed.

[0077] As an example, the service usage address information can be obtained through the service usage time distribution data, so as to determine whether the user's permanent residence area has changed.

[0078] As an example, for users whose resident area preference status changes, we can further analyze whether the network indicators corresponding to the business usage address after the resident area preference change are degraded compared with the network indicators corresponding to the business usage address before the resident area preference change.

[0079] In another exemplary embodiment, the business status includes a business indicator status, and the above step S202 includes:

[0080] Step S2021b: determining the service indicator status according to the service indicator information, the user identification information, and the service behavior status; the service indicator information is the indicator information of the user's service usage;

[0081] Step S2022b: When the business indicator information is inconsistent with the business target indicator information corresponding to the user identification information, or when the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior status, determine that the business indicator status is a changed status.

[0082] As an example, the service indicator information is the indicator information of the service used by the user. For example, the network speed of the user browsing the web is the service indicator information of the user browsing the web.

[0083] For example, if a user's service preferences change, their minimum network requirements may also change. For example, web browsing, short videos, long videos, and gaming all have different network requirements. If current service indicators, such as network conditions, fail to provide a satisfactory user experience after a user's service preferences change, perceptions may not match expectations, leading to complaints.

[0084] As an example, after the user's time preference for using a service terminal changes, the network load will be different due to differences in the number of network users, business behaviors, etc. in different time periods. This will ultimately cause the user's service perception when using the network in the current service usage period to differ from the service expectations when using the network in the user's historical usage period. After the user's time preference changes, if for the same service, the service indicator information of the user's current service usage period, such as the network rate, is lower than the network rate in the user's historical usage period, the service usage perception will be lower than expected, which may easily lead to complaints.

[0085] As an example, after the user's permanent area of ​​use of the service terminal changes, due to certain differences in the geographical environment, network coverage, and network load of different areas, the service indicators such as network speed when the user uses the terminal service in the current area will be different from those in the historical permanent area. This will also cause the perception of service use to be inconsistent with expectations, which is likely to cause complaints.

[0086] Therefore, by analyzing the business target indicator information corresponding to the business indicator information and the user identification information, it can be determined whether the business indicator status has changed. This can be used to analyze whether there is a discrepancy between the user's perception and expectations when using the service, and to predict the probability of user complaints.

[0087] For example, the business indicator information and the business target indicator information corresponding to the user identification information may be analyzed. If the business indicator information and the business target indicator information corresponding to the user identification information are inconsistent, it may be determined that the business indicator status has changed.

[0088] For example, the business indicator information and the business target indicator information corresponding to the business behavior status may be analyzed. If the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior status, it may be determined that the business indicator status has changed.

[0089] For example, the service indicator information for browsing videos includes the network rate. The user's current network rate for browsing videos is 100 mbit / s, while the target indicator information for browsing videos, i.e., the target network rate, is 300 mbit / s. It can be determined that the network rate for browsing videos has changed.

[0090] The process of step S2021b and step S2022b is further described below:

[0091] In an exemplary embodiment, the above step S2021b may include:

[0092] When the business behavior status is a changed status, analyzing the business indicator information after the business behavior information is changed with the business indicator information before the business behavior information is changed to determine the business indicator status;

[0093] In step S2022b, when the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior status, determining that the business indicator status is a changed status may include:

[0094] When the business indicator information after the business behavior information is changed does not meet the business target indicator information corresponding to the current business behavior state, it is determined that the business indicator state is degraded.

[0095] As an example, when the business behavior status is a changed status, the business indicator information after the business behavior information is changed can be analyzed with the business indicator information before the business behavior information is changed. If the business indicator information after the business behavior information is changed is inconsistent with the business indicator information before the business behavior information is changed, it can be determined that the business indicator status has changed.

[0096] As an example, a changed business behavior status may refer to a change in the user's usual business, usual time, or resident area of ​​the terminal. For example, the business behavior that the user spent the most time on before was watching short videos, but recently started playing a certain game, resulting in a change in the business preference status; for example, the time period when the user used the most terminal traffic was from 17:00 to 21:00 in the evening, but recently the time period when the user used the most terminal traffic has changed to 20:00 to 24:00, resulting in a change in the time preference status; for example, the address of the terminal used by the user before was area A, but the address of the terminal used recently is area B, resulting in a change in the resident area preference status. After the user's business behavior status changes, if the current business indicators do not meet the business target indicator information corresponding to the current business behavior status, it may lead to inconsistent user experience of the user using the service.

[0097] As an example, inconsistent user experience can mean a change in the smoothness of a user's current service. This change in smoothness can be due to network issues, resulting in a previously smooth service becoming unsmooth. Alternatively, it can be due to changes in user behavior, with new services placing higher network requirements, resulting in a lower perceived smoothness and inferior user experience compared to previous services.

[0098] For example, for users whose business preferences have changed, we can further analyze whether the business indicators of the business usage type after the change are consistent with the business indicators of the business usage type before the change. If they are inconsistent, for example, the business indicators do not meet the business target indicator information corresponding to the current business preference status, then it can be considered that the business indicators of the current business usage type have deteriorated, which can be used to analyze whether the business indicators have deteriorated after the user changes their business preferences.

[0099] For example, for users whose time preference status has changed, we can further analyze whether the network indicators of the service usage period after the time preference change are consistent with the network indicators of the service usage period before the time preference change. If they are inconsistent, for example, the service indicators do not meet the service target indicator information corresponding to the current time preference status, then it can be considered that the network indicators of the current service usage period have deteriorated. This can be used to analyze whether the service indicators have deteriorated after the user's time preference change. For example, if the network indicators of the current service usage period do not meet the current service smooth use network indicators, then the network indicators have deteriorated.

[0100] For example, for users whose resident area preference status changes, the service indicators of the service usage address after the resident area preference change can be further analyzed to see whether they are consistent with the service indicators of the service usage address before the resident area preference change. If they are inconsistent, for example, if the service indicators do not meet the service target indicator information corresponding to the current resident area preference status, then it can be considered that the service indicators of the current service usage address have deteriorated. This can be used to analyze whether the service indicators have deteriorated after the user changed the resident area preference. For example, if the network indicators after the resident area change cannot meet the requirements for smooth service use, then the network indicators have deteriorated.

[0101] In an exemplary embodiment, the user identification information includes user fee information, and the above step S2021b may include:

[0102] When the service behavior status is unchanged, comparing the current service indicator information with the target service indicator information corresponding to the user tariff information to determine the service indicator status;

[0103] In step S2022b, when the business indicator information is inconsistent with the business target indicator information corresponding to the user identification information, determining that the business indicator state is a changed state includes:

[0104] When the current service indicator information is lower than the service target indicator information corresponding to the user tariff information, it is determined that the service indicator has deteriorated.

[0105] As an example, when the business behavior status has not changed, the user's current business indicator information can be compared with the target business indicator information corresponding to the user's tariff information. If the current business indicator information is lower than the business target indicator information corresponding to the user's tariff information, it can be determined that the business indicator has deteriorated.

[0106] Step S203: predicting the user complaint probability according to the service status.

[0107] As an example, the probability of user complaints can be predicted based on information about whether the business status is consistent or inconsistent.

[0108] In an exemplary embodiment, the above step S203 may include:

[0109] Step S2031, obtaining user historical complaint information;

[0110] Step S2032, calculating the user complaint frequency based on the historical complaint information;

[0111] Step S2033: predicting the user complaint probability based on the service status and the complaint frequency.

[0112] For example, the user complaint frequency may be calculated based on historical complaint information, and the user complaint probability may be predicted based on the calculated user complaint frequency and service status.

[0113] The following is an exemplary description of the application of the complaint prediction method disclosed herein to a user complaint prediction device process:

[0114] The user complaint prediction device may include a real-time data aggregation module, a complaint-prone determination module, and an expected inconsistency determination module.

[0115] (1) Real-time data aggregation module

[0116] It can be used to perform real-time data cleaning and index aggregation for the key KPI indicators of each entity node in the operator's multi-dimensional networking system into basic data for complaint prediction and judgment.

[0117] Basic data can include three parts of information: basic data can include personal O-domain data information, such as {IMSI, CI, terminal brand,...}, personal B-domain data information, such as {gender, user age, user star rating, traffic plan name,...}, and personal O-domain KPI indicator information, such as {webpage effective download rate, video playback duration, video effective download rate, game RTT total delay,...}.

[0118] (2) Complaint-prone judgment module

[0119] It can be used to calculate the frequency or number of user complaints through historical complaint data.

[0120] As an example, the frequency of user complaints can reflect the user's susceptibility to complaints. For example, users who frequently complain can be determined to be users who are prone to complaints.

[0121] (3) Expected inconsistency determination module

[0122] Through the real-time data aggregation module, the use case granularity status identification matrix is ​​constructed as follows:

[0123] [TIME,IMSI1,DIM_O1,DIM_O2,...,DIM_B1,DIM_B2,...,KPI_O1,KPI_O2,...,]

[0124] The expected inconsistency determination module can be used to identify whether the network data actually used by the user is consistent with the network data in the user's tariff package through different time granularities; it can be used to identify whether the user's service preferences have changed through the service usage duration, and further analyze the service indicator status before and after the service change; it can be used to identify whether the user's service usage period has changed through the service usage period, and further analyze the service indicator status before and after the service usage period change; it can be used to identify whether the user's permanent residence area has changed through the service usage address, and further analyze the service indicator status before and after the permanent residence area change.

[0125] The expected inconsistency determination module can also be used to generate an inconsistency matrix based on the above-mentioned change status and business indicator status, wherein the first column in the inconsistency matrix can be time, the second column can be user identification, and each subsequent column represents an inconsistency determination content, where 0 represents consistency and 1 represents inconsistency.

[0126] For example, the inconsistency matrix can be expressed as: [TIME, IMSI, 0, 0, 0, 1, 0, ....].

[0127] The following uses several examples to further illustrate the complaint prediction process:

[0128] Example 1

[0129] User complaint frequency can reflect a user's susceptibility to complaints, while service status can reflect the discrepancy between a user's perception of a service and their expectations, expressed as expectation inconsistency. Expectation inconsistency refers to the degree to which a user's expectations or expectations differ from the situation or outcome of a particular event in a specific scenario.

[0130] User groups can be classified and predicted based on irritability and expectation discrepancy. For example, Figure 3 is a schematic diagram of a complaint prediction model according to an embodiment of the present disclosure. The complaint prediction model can include two dimensions: irritability and expectation discrepancy.

[0131] The entire space can be divided into four areas by dividing the complaint rate and expectation inconsistency rate into high and low levels, as shown in Figure 3:

[0132] (1) Users with low complaint probability and low expectation inconsistency: This type of users are not likely to complain, and their external feedback is consistent with their expectations. This type of users is a zero-complaint group and can be classified as zero-probability complaint.

[0133] (2) Users with high complaint rate and low expectation inconsistency: This type of users are prone to complain. Although external feedback is consistent with expectations, there may be venting of anger complaints due to non-objective reasons, which can be classified as non-objective complaints.

[0134] (3) Users with low complaint probability and high expectation inconsistency: This type of users are not likely to complain. When external feedback is inconsistent with expectations, there is a certain probability of complaint, which can be classified as medium-probability complaints.

[0135] (4) Users with high complaint probability and high expectation inconsistency: This type of users are prone to complain, and there is a high inconsistency between external feedback and expectations. They have a high probability of complaining and can be classified as high-probability complainers.

[0136] Example 2

[0137] User complaint prediction process based on different user behavior differences:

[0138] 1) Obtain basic data.

[0139] Basic data sources can include two parts: 1. Operator user complaint data; 2. Operator O-domain data. O-domain data is used to obtain quantifiable KPI indicators during users' actual network usage.

[0140] 2) Through the user prediction model, the operator's user complaint details data and operator O domain data are aggregated into a basic data table.

[0141] For example, the granular data from August 5, 2023 to August 9, 2023 is obtained. The basic data table after aggregation is as follows, where Table 1 is the basic data table on August 5, 2023, Table 2 is the basic data table on August 9, 2023, and the basic data table from August 6, 2023 to August 8, 2023 is omitted:

[0142] Table 1:

[0143] …

[0144] Table 2:

[0145] 3) Based on the data in the aggregated basic data table and the average complaint cycle calculation formula, calculate the user's average complaint cycle, that is, the complaint frequency. The average complaint cycle calculation formula can be as follows:

[0146] Through historical complaint data, the user complaint statistics are obtained as follows:

[0147] 4) Analysis by the complaint determination module shows that user 163XXXXX001 is a high-frequency complaint user.

[0148] 5) Through the inconsistency determination module, it is identified that the permanent location area of ​​user 163XXXXX001 has changed. A comparative analysis of the regional dimension indicators of areas 747835 and 793145 is performed, and it is determined that area 747835 is overloaded and the average network rate is significantly worse than that of area 793145.

[0149] 6) Through the complaint prediction model, since the permanent residence area of ​​user 163XXXXX001 has changed, and the wireless network indicators in the area after the change have obviously deteriorated; and 163XXXXX001 is a high-frequency complaint user, it is predicted that user 163XXXXX001 has a high probability of complaining.

[0150] Example 3

[0151] User complaint prediction process based on different user behavior differences:

[0152] 1) Obtain basic data.

[0153] Data sources can include three parts: 1. Operator user complaint details; 2. Operator B-domain data; and 3. Operator O-domain data. B-domain data primarily captures basic information such as customer package rates, while O-domain data is used to obtain quantifiable KPI indicators during users' actual network usage.

[0154] Among them, the business can be expanded into five categories: web browsing, long video, short video, game, instant messaging, and voice.

[0155] 2) The user prediction model aggregates the operator's user complaint details and operator O domain data into a basic data table. The granular data for a particular day is obtained, and the aggregated basic data table is shown in Table 3:

[0156] 3) The inconsistency determination module identified user 163XXXXX151 as a 5G plan user with a data plan level of 200. The effective web download rate and video download rate revealed that the user's effective download rate was less than the 1 / 4 quantile of the 5G user indicators for the same plan, and that the user was not a speed-limited user. This is inconsistent with the user's tariff information.

[0157] Through historical complaint data, we obtain the user complaint statistics as shown in Table 4:

[0158] 4) User 163XXXXX151 is a low-frequency complaint user. The complaint prediction model predicts that the user is a medium-probability complaint user.

[0159] The present invention obtains user identification information, business behavior information, and business indicator information, determines the user's business status based on the user identification information, business behavior information, and business indicator information, and predicts the user complaint probability based on the business status. The algorithm logic of the prediction process is visible, and the reason for the user complaint can be obtained based on the predicted user complaint probability and the user, which can be used to guide the implementation of subsequent solutions. It has better practicality and implementation value, and solves the problems in related technologies of complaint prediction methods such as unclear prediction process, poor interpretability of prediction results, inability to guide network optimization personnel to implement the next step of work, and inability to generate practical benefits. It provides strong guarantees for operation and maintenance support and operation analysis of mobile operators, while greatly saving labor costs.

[0160] This embodiment also provides a complaint prediction device for implementing the above-mentioned embodiments and exemplary implementations. Details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.

[0161] FIG4 is a structural block diagram of a complaint prediction device according to an embodiment of the present disclosure. As shown in FIG4 , the device includes:

[0162] Acquisition module 401, configured to acquire user identification information, business behavior information, and business indicator information;

[0163] Determination module 402, configured to determine the user's business status based on the user identification information, the business behavior information, and the business indicator information;

[0164] The prediction module 403 is configured to predict the probability of user complaints based on the service status.

[0165] In an exemplary embodiment, the business behavior information includes current business behavior information and historical business behavior information; the business status includes business behavior status; and the determination module 402 includes:

[0166] A first determination submodule is configured to determine the business behavior status of the user according to the user identification information and the business behavior information corresponding to the user;

[0167] The first determining submodule is configured to determine that the business behavior status is a changed status when the current business behavior information is inconsistent with the historical business behavior information.

[0168] In an exemplary embodiment, the business status includes a business indicator status, and the determining module 402 includes:

[0169] A second determination submodule is configured to determine the service indicator status according to the service indicator information, the user identification information, and the service behavior status; the service indicator information is the indicator information of the user's use of the service;

[0170] The second determination submodule is configured to determine that the business indicator status is a changed status when the business indicator information is inconsistent with the business target indicator information corresponding to the user identification information, or when the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior status.

[0171] In an exemplary embodiment, the service behavior information includes service usage duration information and service usage type information, the service behavior status includes a service preference status, and the first determination submodule includes:

[0172] a first analyzing unit configured to analyze the current service usage duration information and historical service usage duration information of the user for the same type of service based on the user identification information, the service usage duration information, and the service usage type information corresponding to the user, and determine the service preference state of the user;

[0173] The first determining submodule includes:

[0174] The first determining unit is configured to determine that the service preference state is a changed state when an analysis result of the current service usage duration information and the historical service usage duration information of the user for the same type of service are inconsistent.

[0175] In an exemplary embodiment, the service behavior information includes service usage period information, the service behavior status includes a time preference status, and the first determination submodule includes:

[0176] a second analyzing unit configured to analyze the current service usage time period information of the user and the historical service usage time period information according to the user identification information and the service usage time period information corresponding to the user, and determine the time preference state of the user;

[0177] The first determining submodule includes:

[0178] The second determining unit is configured to determine that the time preference state is a changed state when the analysis results of the current service usage period information and the historical service usage period information of the user are inconsistent.

[0179] In an exemplary embodiment, the business behavior information includes business usage address information, the business behavior status includes a resident area preference status, and the first determination submodule includes:

[0180] a third analyzing unit configured to analyze the current service usage address information and the historical service usage address information of the user according to the user identification information and the service usage address information corresponding to the user, and determine the resident area preference state of the user;

[0181] The first determining submodule includes:

[0182] The third determining unit is configured to determine that the resident zone preference state of the user is a changed state when the current service usage address information of the user is inconsistent with the historical service usage address information.

[0183] In an exemplary embodiment, the second determination submodule includes:

[0184] a fourth analyzing unit configured to, when the business behavior status is a changed status, analyze the business indicator information after the business behavior information is changed with the business indicator information before the business behavior information is changed, and determine the business indicator status;

[0185] The second determining submodule includes:

[0186] The fourth determining unit is configured to determine that the business indicator state is a changed state when the business indicator information after the business behavior information is changed is inconsistent with the business indicator information before the business behavior information is changed.

[0187] In an exemplary embodiment, the user identification information includes user fee information, and the second determination submodule includes:

[0188] a comparing unit configured to compare the current service indicator information with the target service indicator information corresponding to the user tariff information to determine the service indicator status when the service behavior status is unchanged;

[0189] The second determining submodule includes:

[0190] The fifth determining unit is configured to determine that the service indicator has deteriorated when the current service indicator information is lower than the service target indicator information corresponding to the user tariff information.

[0191] In an exemplary embodiment, the prediction module 403 includes:

[0192] Get submodule, set to get user historical complaint information;

[0193] A calculation submodule, configured to calculate the user complaint frequency based on the historical complaint information;

[0194] The prediction submodule is configured to predict the probability of user complaints based on the service status and the complaint frequency.

[0195] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0196] An embodiment of the present disclosure further provides a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps of any one of the above method embodiments when run.

[0197] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0198] An embodiment of the present disclosure further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0199] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0200] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail here.

[0201] Obviously, those skilled in the art should understand that the modules or steps of the present disclosure described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed across a network composed of multiple computing devices, they can be implemented using program code executable by the computing device, and thus, they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be performed in a different order than herein, or they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present disclosure is not limited to any particular combination of hardware and software.

[0202] The foregoing is merely an exemplary embodiment of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art will readily appreciate that the present disclosure is susceptible to various modifications and variations. Any modifications, equivalent substitutions, improvements, and the like made within the principles of the present disclosure shall be included within the scope of protection of the present disclosure.

Claims

1. A complaint prediction method, comprising: Obtain user identification information, business behavior information, and business indicator information; Determining the user's business status according to the user identification information, the business behavior information, and the business indicator information; The user complaint probability is predicted according to the service status.

2. The method according to claim 1, wherein: The business behavior information includes current business behavior information and historical business behavior information; the business status includes business behavior status; Determining the user's business status according to the user identification information, business behavior information, and business indicator information includes: Determining the business behavior status of the user according to the user identification information and the business behavior information corresponding to the user; When the current business behavior information is inconsistent with the historical business behavior information, it is determined that the business behavior state is a changed state.

3. The method according to claim 2, wherein: The business status includes a business indicator status, and determining the business status of the user according to the user identification information, the business behavior information, and the business indicator information includes: Determining the service indicator status according to the service indicator information, the user identification information, and the service behavior status; the service indicator information is the indicator information of the user's use of the service; When the business indicator information is inconsistent with the business target indicator information corresponding to the user identification information, or when the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior status, it is determined that the business indicator status is a changed status.

4. The method according to claim 2, wherein: The service behavior information includes service usage duration information and service usage type information, the service behavior status includes service preference status, and the determining the service behavior status of the user according to the user identification information corresponding to the user and the service behavior information includes: According to the user identification information, the service usage duration information, and the service usage type information corresponding to the user, the current service usage duration information and the historical service usage duration information of the user for the same type of service are analyzed to determine the service preference state of the user; When the current business behavior information is inconsistent with the historical business behavior information, determining that the business behavior state is a changed state includes: When the analysis results of the current service usage duration information and the historical service usage duration information of the user for the same type of service are inconsistent, it is determined that the service preference state is a changed state.

5. The method according to claim 2, wherein: The business behavior information includes business usage period information, the business behavior status includes time preference status, and the determining the business behavior status of the user according to the user identification information and the business behavior information corresponding to the user includes: According to the user identification information and the service usage period information corresponding to the user, analyzing the current service usage period information of the user and the historical service usage period information, and determining the time preference state of the user; When the current business behavior information is inconsistent with the historical business behavior information, determining that the business behavior state is a changed state includes: When the analysis results of the current service usage period information and the historical service usage period information of the user are inconsistent, it is determined that the time preference state is a changed state.

6. The method according to claim 2, wherein: The business behavior information includes business usage address information, the business behavior status includes a resident area preference status, and the determining the business behavior status of the user according to the user identification information and the business behavior information corresponding to the user includes: According to the user identification information and the service usage address information corresponding to the user, analyzing the current service usage address information and the historical service usage address information of the user, and determining the resident area preference state of the user; When the analysis results of the current business behavior information and the historical business behavior information are inconsistent, determining that the business behavior state is a changed state includes: When the current service usage address information of the user is inconsistent with the historical service usage address information, it is determined that the resident area preference state of the user is a changed state.

7. The method according to claim 3, wherein: Determining the business indicator status according to the business indicator information, the user identification information, and the business behavior status includes: When the business behavior status is a changed status, analyzing the business indicator information after the business behavior information is changed with the business indicator information before the business behavior information is changed, and determining the business indicator status; When the business indicator information is inconsistent with the business target indicator information corresponding to the business behavior state, determining that the business indicator state is a changed state includes: When the business indicator information after the business behavior information is changed does not satisfy the business target indicator information corresponding to the current business behavior state, it is determined that the business indicator state is degraded.

8. The method according to claim 3, wherein: The user identification information includes user fee information, and the determining the service indicator status according to the service indicator information, the user identification information, and the service behavior status includes: When the service behavior state is unchanged, comparing the current service indicator information with the target service indicator information corresponding to the user tariff information to determine the service indicator state; When the business indicator information is inconsistent with the business target indicator information corresponding to the user identification information, determining that the business indicator state is a changed state includes: When the current service indicator information is lower than the service target indicator information corresponding to the user tariff information, it is determined that the service indicator has deteriorated.

9. The method according to claim 1, wherein: The predicting the user complaint probability according to the service status includes: Obtain user historical complaint information; Calculating the user complaint frequency based on the historical complaint information; The user complaint probability is predicted according to the service status and the complaint frequency.

10. A complaint prediction device, comprising: An acquisition module, configured to acquire user identification information, business behavior information, and business indicator information; A determination module, configured to determine the service status of the user according to the user identification information, the service behavior information, and the service indicator information; The prediction module is configured to predict the probability of user complaints based on the business status.

11. A computer-readable storage medium having a computer program stored therein, wherein: When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 9 are implemented.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in any one of claims 1 to 9 when executing the computer program.

Images