Business complaint data analysis method, device, equipment and medium

Abstract

The application relates to the technical field of data analysis, and discloses a business complaint data analysis method, device, equipment and medium.The method comprises the following steps: identifying a complaint customer identity in a historical complaint data set of a target business, dividing the target complaint data set into a historical complaint customer set and a new complaint customer set according to the complaint customer identity and a preset time reference point; constructing a repeated complaint probability function of the historical complaint customer set, and calculating a first total amount of complaints according to the repeated complaint probability function; extracting a complaint reference amount and a complaint event influence factor of the new complaint customer set; correcting and calculating the complaint reference amount according to the complaint event influence factor, so as to obtain a second total amount of complaints; and integrating the first total amount of complaints and the second total amount of complaints into an expected total amount of complaints; determining a target amount of complaints based on the expected total amount of complaints, and performing visual display.The application can be applied to a financial technology, medical health and other business system platform, and can improve the accuracy of business complaint data analysis.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a method, apparatus, equipment and medium for analyzing business complaint data. Background Technology

[0002] As business scale continues to expand, business complaint data (e.g., data on the number of complaints related to a business) has become one of the core indicators for measuring service quality and operational health. In existing technologies, business complaint data analysis methods mainly rely on simple statistics of historical data or manual experience to analyze the trend of complaint volume data, lacking a scientific and dynamic prediction mechanism.

[0003] Specifically, existing methods typically treat business complaint data as a holistic indicator without effectively breaking down the sources of complaints. For example, they fail to distinguish the characteristics and trends of repeat customers and new customers who file complaints. This results in a lack of specificity in complaint trend analysis, making it difficult to reflect actual business risks and changes in customer complaint behavior. It can easily lead to biased predictive analysis and is difficult to support refined decision-making on complaint volume control.

[0004] For example, in the healthcare context, business complaint data analysis, such as the analysis of patient complaint volume, is a key indicator for evaluating hospital service quality. Existing technologies typically combine the total number of complaints from outpatient, inpatient, and online consultation channels, failing to distinguish between repeat and first-time complainants. Repeat complainants exhibit highly predictable complaint behavior, while the number of new complainants is significantly influenced by external factors such as public health emergencies and policy adjustments. This combined analysis leads to substantial prediction bias, making it difficult to support refined service quality control in hospitals.

[0005] For example, in fintech scenarios, business complaint data analysis, such as the analysis of customer complaint volume, is directly related to platform compliance and user trust. Existing methods often aggregate complaint data by product line, failing to effectively separate existing and new complaining customers. In other words, current technology lacks differentiated modeling and dynamic correction mechanisms for these two types of customers, resulting in a lack of scientific basis for setting complaint volume control thresholds and making it difficult to achieve proactive risk warnings and refined operations.

[0006] Therefore, improving the accuracy of business complaint data analysis is a pressing technical problem that needs to be solved. Summary of the Invention

[0007] This invention provides a method, apparatus, equipment, and medium for analyzing business complaint data, the main purpose of which is to solve the problem of low accuracy in business complaint data analysis.

[0008] Firstly, to achieve the above objectives, the present invention provides a business complaint data analysis method, comprising: Obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset; Identify the identity of complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaint customer identity and a preset time reference point. Construct a re-complaint probability function for each historical complainant in the historical complainant set, and calculate the first total number of complaints from the historical complainant in a future preset time period based on the re-complaint probability function; Extract the baseline complaint volume for each new customer in the newly added complaint customer set, and collect various complaint event influencing factors for the newly added complaint customer set within the preset time period; The baseline complaint volume is adjusted and calculated based on various factors affecting the complaint events to obtain the second total complaint volume of the newly added customer set within the preset time period. The total number of the first complaint and the total number of the second complaint are combined to generate the expected total number of complaints within the preset time period; The target complaint volume for the target business is determined based on the expected total complaint volume, and the expected total complaint volume and the target complaint volume are visualized.

[0009] Secondly, the present invention also provides a business complaint data analysis device, comprising: The structured processing module is used to obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset. The customer set segmentation module is used to identify the identity of the complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaining customer identity and a preset time reference point. The first complaint volume calculation module is used to construct the re-complaint probability function for each historical complaint customer in the historical complaint customer set, and calculate the first total number of complaints from the historical complaint customers in a future preset time period based on the re-complaint probability function. The complaint factor extraction module is used to extract the complaint baseline quantity for each new customer in the new complaint customer set, and to collect various complaint event influencing factors of the new complaint customer set within the preset time period. The second complaint volume calculation module is used to correct the complaint baseline volume based on various complaint event influencing factors to obtain the second total number of complaints for the newly added complaint customer set within the preset time period. The expected complaint volume analysis module is used to summarize the first total complaint volume and the second total complaint volume to generate the expected total complaint volume within the preset time period; The visualization module is used to determine the target complaint volume for the target business based on the expected total complaint volume, and to visualize the expected total complaint volume and the target complaint volume.

[0010] Thirdly, the present invention also provides an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the business complaint data analysis method described above.

[0011] Fourthly, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the business complaint data analysis method described above.

[0012] In this embodiment of the invention, multi-source, unstructured historical complaint data is transformed into a highly cohesive, low-redundancy target dataset, significantly reducing resource overhead and memory consumption in subsequent computational stages. Based on customer identification and time reference points, hierarchical segmentation of the dataset enables dynamic partitioning and pruning, allowing for the creation of a long-term cached view of the historical set to support incremental iterative computation. Simultaneously, real-time streaming processing of new data avoids the computational waste associated with full-scale scanning. By constructing a personalized re-complaint probability function for each historical customer, the heterogeneity and temporal dependence of complaint behavior are captured, significantly improving prediction accuracy compared to the global mean model. For new customers, a separate modeling architecture is introduced, combining a baseline complaint volume with dynamic event influencing factors, achieving low-latency prediction and high accuracy of online business complaint data while maintaining precision. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of an application environment for a business complaint data analysis method according to an embodiment of the present invention. Figure 2 A flowchart illustrating a business complaint data analysis method provided in an embodiment of the present invention; Figure 3 A functional block diagram of a business complaint data analysis device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device for implementing a business complaint data analysis method according to an embodiment of the present invention; Figure 5 This is another structural schematic diagram of an electronic device for implementing a business complaint data analysis method according to an embodiment of the present invention.

[0015] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0016] It should be noted that in the technical solutions disclosed in this invention, the acquisition of user information (personal image data (e.g., facial videos or pictures, facial feature videos or pictures, etc.) and personal privacy information (e.g., name, ID number, occupation, address, etc.)) is all completed with the user's knowledge and consent, and the acquisition of the relevant user information is legal and compliant.

[0017] To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0018] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0019] This application provides a business complaint data analysis method. The execution subject of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the device provided in this application: a server, a terminal, or other similar device. In other words, the business complaint data analysis method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] The business complaint data analysis method of this invention can be applied to, for example... Figure 1 In this application environment, the client communicates with the server via a network. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will now be described in detail through specific embodiments.

[0021] Reference Figure 2 The diagram shown is a flowchart illustrating a business complaint data analysis method according to an embodiment of the present invention. In this embodiment, the business complaint data analysis method includes: S1. Obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset.

[0022] In this embodiment of the invention, the historical complaint dataset refers to a collection of multi-source heterogeneous data originally collected from various complaint channels of the target business (such as customer service tickets, online feedback, emails, social media comments, etc.) without systematic processing. This dataset typically contains unstructured text (complaint content descriptions), semi-structured logs (timestamps, channel identifiers), and structured fields (customer ID, product category, etc.), and may contain data quality issues such as missing values, duplicate records, and inconsistent formats.

[0023] In this embodiment of the invention, all complaint-related records generated within a historical timeframe for a specific business line are extracted in batches from the complaint management module, customer service ticket platform, online feedback channels, and historical archive database of the business system, using the business identifier of the target business as the filtering condition. Records from different data sources are uniformly accessed, eliminating data barriers caused by differences in storage formats, field naming, and encoding methods between systems. This aggregates complaint information scattered across various business systems into a single original historical complaint dataset containing a complete set of complaint events, ensuring comprehensive and complete data coverage for subsequent processing.

[0024] The process involved multiple rounds of cleaning the original historical complaint dataset. First, missing value detection identified and removed invalid records with empty key fields. Then, duplicate complaint entries caused by repeated system pushes were removed through duplicate detection. Next, unstructured complaint description text was segmented and semantically annotated, transforming it into machine-readable structured labels. Finally, the cleaned records underwent field alignment and type conversion according to a unified data standard. This reshaped the originally inconsistent and low-quality raw data into a target complaint dataset where each record has a unified field structure, data type, and coding standard, meeting the stringent consistency and standardization requirements of subsequent automated analysis processes.

[0025] For example, in a healthcare scenario, all patient complaint records from a designated department within a specific historical period are extracted from the hospital complaint management system, online consultation platform, and satisfaction survey module, and aggregated into a historical complaint dataset. During structured processing, the patient's complaint text is processed using medical named entity recognition technology to extract symptom keywords. The reasons for the complaint are mapped to unified medical dispute classification tags, the consultation time is uniformly converted to a standard time format, and the patient's name and medical card number are anonymized and normalized. Finally, a target complaint dataset is generated, with each record containing unified fields such as department, complaint category, severity level, timestamp, and anonymized identity identifier.

[0026] For example, in a fintech scenario, all customer complaint records for a specific wealth management product line within a given historical period are extracted from the bank's customer complaint system, wealth management product feedback platform, and regulatory transfer order database, and aggregated into a historical complaint dataset. During structured processing, key information such as product type, transaction amount, and violations are extracted from the customer complaint description using financial named entity recognition technology. The reasons for the complaints are mapped to unified financial consumer dispute classification tags, transaction times are standardized to a standard timestamp format, and customer names and account identifiers are encrypted, desensitized, and standardized in format. Finally, a target complaint dataset is generated, with each record containing unified fields such as product line, dispute type, risk level, timestamp, and desensitized identity identifier.

[0027] S2. Identify the identity of the complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaint customer identity and a preset time reference point.

[0028] In this embodiment of the invention, the historical complaint customer set consists of customers who have complaint records before the time reference point, and the newly added complaint customer set consists of customers whose complaint records appear for the first time at the time reference point.

[0029] In this embodiment of the invention, identifying the complainant's identity in the target complaint dataset includes: Obtain each complaint record in the target complaint dataset, and extract at least one preset identity association field from the complaint record; The identity association field is formatted and standardized to obtain the target identity association field; Each of the target identity association fields is matched with the corresponding field in the preset customer profile database to obtain the confidence value of each target identity association field that is successfully matched; The target identity association field corresponding to the matching result whose confidence value is greater than or equal to the preset confidence threshold is determined as the complainant customer identity identifier in the complaint record.

[0030] In this embodiment of the invention, each complaint record in the target complaint dataset is scanned one by one. Complaint records are typically stored in text or semi-structured format, containing information such as name, mobile phone number, ID card number, order number, and email address. Using preset field rule templates, regular expression matching and named entity recognition technology in natural language processing are employed to locate and extract customer-identity-related fields from the unstructured or semi-structured text content. These fields are referred to as identity-related fields.

[0031] In this process, because the same type of identity information may be written differently in different complaint records—for example, mobile phone numbers may have area codes or not, some may be separated by hyphens or spaces, names may have full names or abbreviations, and ID card numbers may contain uppercase and lowercase letters—character cleaning is performed on each extracted identity-related field. This involves removing redundant spaces, special symbols, and converting between full-width and half-width characters. For mobile phone numbers, international area codes are added or removed. For names, traditional and simplified characters are converted and spaces are removed. For ID card numbers, uppercase and lowercase are standardized and the number of digits is verified. After this step, the target identity-related fields are uniformly formatted, standardized, and can be directly used for comparison, eliminating matching obstacles caused by different writing habits.

[0032] Specifically, each standardized target identity association field is compared with the corresponding field pre-stored in the customer profile database using a fuzzy matching algorithm, such as string similarity calculation based on edit distance. Each comparison outputs a confidence value, which reflects the degree of similarity or match between the target identity association field and the corresponding field of a customer record in the database. The higher the confidence value, the more likely the two are to point to the same person. Multiple confidence values ​​are calculated independently for each field, representing the reliability of different identity clues pointing to the same customer.

[0033] Furthermore, the confidence value of each target identity association field is compared with this threshold. Only when the confidence value reaches or exceeds the threshold is the matching result of that field considered reliable, and the corresponding target identity association field is identified as the complainant's identity identifier in the complaint record. Matching results with a confidence value below the threshold are considered noise or misjudgments and are discarded directly. The purpose of this is to filter out unreliable matches while ensuring recognition accuracy, ultimately outputting a verified, highly reliable customer identity identifier for subsequent complaint processing and customer service procedures.

[0034] For example, in a healthcare scenario, patient registration numbers are extracted from the target complaint dataset as identifiers for complaining customers. Using the date the hospital launched its complaint warning system as a reference point, patient registration numbers with complaint records prior to that date are categorized into a historical complaint customer set, including chronic disease patients who have repeatedly visited and complained. Patient registration numbers with their first complaint record after that date are categorized into a new complaint customer set, including initial visit patients who complained for the first time due to a poor experience during a single visit. The two customer sets are then used in subsequent predictions, with the re-complaint probability model and the influence factor correction model applied respectively.

[0035] For example, in a fintech scenario, encrypted account identifiers are extracted from the target complaint dataset to identify complaining customers. Using the effective date of the latest consumer rights protection guidelines issued by the regulatory authorities as a reference point, customer accounts with complaint records prior to that date are categorized into the historical complaint customer set, including existing customers who have repeatedly purchased wealth management products and filed complaints. Customer accounts with their first complaint record after that date are categorized into the new complaint customer set, including newly opened accounts that filed complaints for the first time due to losses from their first fund purchase. These two customer sets are then processed through different complaint volume prediction processes to accommodate the differences in their complaint behavior characteristics.

[0036] In this embodiment of the invention, by matching identity identifiers and cutting timestamps, the spatiotemporal trajectory of customer complaint behavior can be efficiently constructed, achieving low redundancy and high cohesion storage of the dataset, and significantly reducing the resource overhead and memory usage of subsequent analysis and computation.

[0037] S3. Construct a re-complaint probability function for each historical complainant in the historical complainant set, and calculate the first total number of complaints from the historical complainant in a future preset time period based on the re-complaint probability function.

[0038] In this embodiment of the invention, the re-complaint probability function refers to a mathematical model constructed for each historical complaint customer, used to describe the likelihood of them filing a complaint again at a specific point in time or within a time interval in the future. This function is typically personalized based on characteristics such as the customer's historical complaint frequency, time interval, and complaint channels (e.g., using logistic regression, survival analysis, or Poisson process), and the output value is a probability distribution between 0 and 1. The first total complaint amount is an estimated number of complaints obtained by summing or integrating the re-complaint probabilities of each historical complaint customer within a preset future time period. This value is not a simple historical average, but an expected value dynamically calculated through the probability function, reflecting the potential number of complaints from this customer group in the future.

[0039] In this embodiment of the invention, constructing the re-suit probability function for each historical complainant in the historical complainant set includes: Extract the complaint time series of each historical complaint customer from each complaint record in the historical complaint customer set, and calculate the time interval between two adjacent complaint records based on the complaint time series to obtain the complaint interval sequence; Using the preset maximum likelihood estimation method, the parameters of at least two preset candidate probability distribution models are fitted according to the complaint interval sequence to obtain the fitted probability distribution model. Calculate the goodness-of-fit index between each of the fitted probability distribution models and the complaint interval sequence; Based on the goodness-of-fit index, the optimal probability distribution model is selected from the fitted probability distribution models; The cumulative distribution function corresponding to the optimal probability distribution model is determined as the re-suit probability function for the historical complaint customers.

[0040] In this embodiment of the invention, complaint records are retrieved one by one from the historical customer complaint set. The complaint events of each customer are sorted chronologically to form an ordered complaint time sequence. Then, the time span is calculated by subtracting the timestamps of two adjacent complaints to obtain the complaint interval sequence for that customer. This sequence reflects the rhythmic characteristics of repeated complaints by customers and serves as the basic data input for subsequent probability modeling.

[0041] Specifically, several candidate probability distribution models are preset, and the complaint interval sequence is used as the observation sample. Parameter estimation is performed on each candidate model. The core idea of ​​maximum likelihood estimation is to find a set of parameters that maximizes the probability of observing the current complaint interval sequence. The parameters of each candidate model are adjusted through an iterative optimization algorithm so that the theoretical distribution generated by the model is as close as possible to the actual observed interval data, thereby obtaining the fitted probability distribution model corresponding to each candidate model.

[0042] In detail, statistical measures that can measure the degree of deviation between the theoretical distribution and the actual data are selected, and the goodness-of-fit index between each fitting probability distribution model and the complaint interval sequence is calculated. This index quantifies the model's ability to interpret real data. The better the index value, the higher the degree of agreement between the model and the actual interval data, providing an objective basis for subsequent model selection.

[0043] Furthermore, the goodness-of-fit indices of each fitted probability distribution model are compared horizontally, and the models are ranked according to their merits. The model with the best goodness-of-fit is selected as the optimal probability distribution model. This step essentially involves selecting the distribution type that best describes the customer's retrial interval pattern from multiple candidate models, ensuring the reliability of the subsequent probability function.

[0044] Furthermore, the cumulative distribution function corresponding to the optimal probability distribution model is directly defined as the re-complaint probability function of the historical complaint customer. The physical meaning of the cumulative distribution function is: the probability that the customer will complain again before any given time point. This function maps the time variable to the re-complaint probability value. The longer the time, the closer the probability is to a certain value, thus completely depicting the evolution of the customer's future re-complaint probability over time, providing a quantitative basis for complaint early warning and service strategies.

[0045] In this embodiment of the invention, calculating the total number of complaints from the historical complaint customers within a preset future time period based on the re-complaint probability function includes: Obtain the duration of a future preset time period, and construct a prediction time window based on the duration of the period; The prediction time window is divided into multiple consecutive and non-overlapping unit time intervals; The probability of a historical complaint customer filing at least one complaint within the unit time interval is calculated based on the re-complaint probability function. The interval complaint probabilities corresponding to all the unit time intervals are summed, and the sum is rounded up to obtain the first total number of complaints from the historical complaint customers in the future preset time period.

[0046] In this embodiment of the invention, the duration of a preset future time period specified by the user is obtained. This duration defines the boundary of the time range to be predicted. Starting from the current moment and ending with this duration, a continuous prediction time window is constructed. This window limits the time domain range of all subsequent probability calculations, ensuring that the repetition probability function is only called within this window, thereby truncating the probability distribution on the infinite time axis into a prediction task within a finite interval.

[0047] Specifically, the prediction time window is divided into multiple continuous and non-overlapping unit time intervals. The essence of this operation is to transform the continuous time domain into a discrete time series, so that the repetition probability function can be evaluated independently in each small interval. Discretization transforms the probability calculation problem that originally required integration over continuous time into an accumulation problem of summing the probabilities of each small interval, which greatly reduces the computational complexity and improves the feasibility of engineering.

[0048] The process involves substituting the start and end times of each unit time interval into the re-complaint probability function, i.e., the customer's cumulative distribution function, to obtain the cumulative probability values ​​corresponding to the start and end points of the interval. Subtracting the cumulative probability at the start point from the cumulative probability at the end point yields the interval complaint probability for that customer within that unit time interval. This difference represents the conditional probability that the customer has not filed a complaint before the start point of the interval but has filed a complaint before the end point, precisely characterizing the likelihood of a complaint event falling within that small interval.

[0049] Furthermore, by summing the interval complaint probabilities corresponding to all unit time intervals, the resulting sum is mathematically equivalent to the expected number of complaints the customer will make within the entire prediction time window. Since the number of complaints must be an integer, this expected value is rounded up to the nearest integer not less than the expected value, serving as the initial total number of complaints for this historically complaining customer in the future preset time period. This result represents a conservative estimate of the customer's future complaint frequency, ensuring that the early warning strategy does not overlook potential risks due to rounding down.

[0050] In this embodiment of the invention, by constructing a re-complaint probability function for each historical complaint customer and predicting their future total number of complaints, the heterogeneity of customer complaint behavior can be captured, significantly improving prediction accuracy compared to the global average. Secondly, when using this function to estimate the total amount, incremental calculation or approximate aggregation techniques can be used to avoid repeated scanning of the full historical data, achieving lightweight and low-latency online prediction.

[0051] S4. Extract the baseline complaint volume for each new customer in the newly added complaint customer set, and collect various complaint event influencing factors of the newly added complaint customer set within the preset time period.

[0052] In this embodiment of the invention, the complaint baseline value refers to the initial complaint expectation value set individually for each new customer who makes a complaint. It represents the basic complaint tendency under the influence of no external special events. The baseline value is usually assigned a value based on static characteristics such as the specific reasons for the customer's first complaint, the complaint channel, the product category, and the customer's basic attributes (such as membership level and consumption history). For example, it can be set to 1 or standardized according to the average complaint rate of similar customer groups.

[0053] The aforementioned complaint event impact factors refer to various dynamic variables collected from newly added complaint customer sets, used to quantify the amplifying or inhibiting effects of various external or internal events on customer complaint behavior. Common factors include: online public opinion indices (such as Baidu Index, Weibo mentions), external policy events, and external calendar events.

[0054] In this embodiment of the invention, the initial complaint record of each customer is retrieved one by one from the newly added complaint customer set. Using the time of the customer's first complaint as the anchor point, the historical interaction data before the customer became a complaining customer is traced back, including the number of inquiries, the frequency of returns and exchanges, the number of service tickets, etc. These pre-complaint interactions are weighted and aggregated, and transformed into a quantitative value that reflects the degree of dissatisfaction shown by the customer before formally filing a complaint. This value is the complaint baseline quantity. The complaint baseline quantity is essentially a baseline estimate of the potential complaint intention of new customers. It does not rely on any external influencing factors, but is calculated solely based on the customer's own historical behavior patterns, providing a clean starting point for subsequent adjustments to influencing factors.

[0055] This invention synchronously collects various influencing factors that drive complaint behavior from multiple external data sources within a preset time period. These factors cover macro-level factors such as industry policy changes, market sentiment, and competitor service events; meso-level factors such as seasonal climate characteristics, holiday arrangements, and promotional activity schedules; and micro-level factors such as product failure rates, logistics delay rates, and customer service response times. All collected factors are uniformly aligned according to the time dimension, forming a multi-dimensional set of influencing factors covering the entire preset time period. Each factor is standardized into a quantifiable intensity indicator, collectively forming a complete factor pool for subsequent correction calculations of the complaint baseline, ensuring that the correction process covers all internal and external driving factors that may affect new customer complaint behavior.

[0056] For example, in a healthcare scenario, a baseline complaint volume is extracted from each newly complaining patient. This volume is then weighted and aggregated based on the patient's prior behaviors, such as the number of outpatient consultations, the number of canceled appointments, and the duration of doctor-patient communication before the initial complaint, to obtain a baseline complaint volume reflecting the patient's potential level of dissatisfaction. Subsequently, various influencing factors during this period are collected, including factors such as the surge in outpatient visits during seasonal influenza outbreaks, the shortage of on-duty doctors during holidays, the shortage of specific drugs, and changes in the medical process due to adjustments in healthcare reform policies. This forms a multi-dimensional influencing factor pool covering the entire prediction period, which is used for subsequent personalized adjustments to the baseline complaint volume.

[0057] For example, in a fintech scenario, a baseline complaint volume is extracted from each newly complaining customer. This volume is weighted and aggregated based on the customer's prior behavior, such as account activity, frequency of product inquiries, number of failed transactions, and customer service connection wait times, to obtain a baseline complaint volume reflecting their potential complaint intentions. Subsequently, various influencing factors during this period are collected, including customer anxiety due to significant market volatility, redemption pressure caused by concentrated maturity of wealth management products, business adjustments due to tightened regulatory policies, and customer confidence shaken by the spread of negative public opinion about competitors. This forms a multi-dimensional influencing factor pool covering the entire prediction period, which is used for subsequent personalized adjustments to the complaint baseline volume.

[0058] In this embodiment of the invention, by extracting an independent complaint baseline for each newly added customer, a personalized baseline model can be established, avoiding the view that all new customers are a homogeneous group, thereby improving the granularity and accuracy of the prediction. By collecting various factors affecting complaint events (such as response time, public opinion heat, promotional activities, etc.) as dynamic feature variables, the algorithm model can perceive the driving effect of changes in the external environment on customer behavior, realizing the upgrade from static estimation to dynamic adaptive prediction, and improving the accuracy of business complaint analysis.

[0059] S5. Based on various factors affecting the complaint events, the baseline complaint volume is adjusted and calculated to obtain the second total number of complaints from the newly added customer set within the preset time period.

[0060] In this embodiment of the invention, the step of correcting the baseline complaint volume based on various influencing factors of the complaint event to obtain the second total number of complaints for the newly added customer set within the preset time period includes: Obtain the time series of the influence intensity of each of the complaint event's influence factors over time within the preset time period, and align the time series of influence intensity on the time axis to generate an influence factor matrix; Match the set of sensitivity coefficients of the newly added complaining customer to each of the factors affecting the complaint event from the preset factor sensitivity configuration table; Each sensitivity coefficient in the sensitivity coefficient set is multiplied point by point with the corresponding influence intensity time series in the influence factor matrix to obtain the total correction coefficient for the newly added complaining customers; Multiplying the baseline number of complaints by the total correction factor yields the second total number of complaints from the newly added customer set within the preset time period.

[0061] In this embodiment of the invention, the time series of the influence intensity of each complaint event's influencing factor changes over time within a preset time period are obtained. These series reflect the strength of each factor at different time points. Since the time sampling frequency and start and end times of each factor may be different, a unified time alignment operation is performed on all influence intensity time series based on the time axis of the preset time period. They are matched and filled one by one according to the timestamp to generate an influence factor matrix with time as the row and influence factor as the column. This matrix completely depicts the spatiotemporal distribution of all influence factors throughout the entire prediction period, providing a unified data carrier for subsequent weighted correction.

[0062] Specifically, the sensitivity coefficients of newly added complaining customers to various influencing factors are retrieved and matched from a pre-maintained factor sensitivity configuration table based on their attribute characteristics. These sensitivity coefficients reflect a specific customer's responsiveness to a particular influencing factor; different customers exhibit varying sensitivities to the same factor due to differences in consumption habits, complaint history, and service preferences. Using the customer identifier as a key, a precise lookup is performed in the configuration table to extract all corresponding sensitivity coefficients for that customer, forming a set of sensitivity coefficients that correspond one-to-one with each influencing factor. This set serves as the weighting source for subsequent weighted calculations.

[0063] Specifically, each sensitivity coefficient in the sensitivity coefficient set is multiplied at each time point by the time series of the influence intensity in the corresponding column of the influence factor matrix. Essentially, this operation scales the actual influence intensity of each influence factor at each time point using the customer's sensitivity to that factor, thus obtaining a corrected influence intensity time series after personalized weighting. Subsequently, the weighted results of all factors at all time points are accumulated and aggregated to obtain a comprehensive correction coefficient that reflects the combined effect of all influence factors on the customer throughout the entire prediction period. This coefficient is a global correction value that integrates both time and factor dimensions.

[0064] Furthermore, the predetermined baseline complaint volume is multiplied by the total correction factor. The baseline complaint volume represents the theoretical number of complaints from the new customer set without external influencing factors, while the total correction factor quantifies the combined amplification or reduction effect of various influencing factors on the customer set's complaint behavior. After multiplying, the baseline volume is proportionally adjusted by the total correction factor, ultimately yielding the second total number of complaints from the new customer set within a preset time period. This result is the actual predicted value after adding corrections for influencing factors to the baseline prediction, making it more closely aligned with real-world business scenarios than the baseline volume alone.

[0065] In this embodiment of the invention, the baseline quantity of complaints is dynamically corrected by various factors affecting complaint events, thereby achieving adaptive modeling of new customer complaint behavior and avoiding systematic biases caused by static baseline quantities when the environment changes. Through interpretability analysis of factor weights, the key events that have the most significant impact on the total number of complaints (such as response delays or public opinion outbreaks) can be automatically identified, providing a quantitative priority ranking for operational intervention strategies and promoting the intelligent closed-loop evolution of complaint prediction from "total quantity estimation" to "attribution-driven".

[0066] S6. Summarize the first total number of complaints and the second total number of complaints to generate the expected total number of complaints within the preset time period.

[0067] In this embodiment of the invention, the step of summing the first total number of complaints and the second total number of complaints to generate the expected total number of complaints within the preset time period includes: The initial total number of complaints is obtained by summing the first total number of complaints and the second total number of complaints. Obtain seasonal and holiday characteristic information contained within the preset time period, and make a first correction to the initial total number of complaints based on the seasonal and holiday characteristic information to obtain the corrected total number of complaints; Determine whether a preset special event is included within the preset time period; If included, the historical complaint volume of the special event during the same historical time period is extracted from the preset historical complaint database, and the deviation value between the historical complaint volume and the preset benchmark complaint volume is calculated. The deviation value is added to the corrected total number of complaints to obtain the expected total number of complaints within the preset time period.

[0068] In this embodiment of the invention, the first total number of complaints calculated based on the probability function of repeat complaints from historical complaining customers is directly added to the second total number of complaints corrected based on the influence factors of newly complaining customers. Essentially, this operation merges the predicted complaint volumes of the two customer groups. The first total number of complaints reflects the complaint expectations of existing customers due to repeat complaint inertia, while the second total number of complaints reflects the complaint expectations of new customers driven by external influencing factors. Adding the two together yields the original expected total number of complaints without any scenario correction, which serves as the baseline input for subsequent corrections.

[0069] This process involves retrieving seasonal and holiday characteristics from a pre-defined feature library for the current preset time period. These characteristics reflect the inherent fluctuations in complaint behavior across different seasons and holidays. The initial total complaint volume is then proportionally adjusted based on this information—for example, increasing it during peak complaint periods and decreasing it during off-peak periods—resulting in a revised total complaint volume. This revision eliminates the bias inherent in purely probabilistic summations that do not consider periodic fluctuations, making the predicted value more closely aligned with the actual seasonal rhythm of business operations.

[0070] Specifically, the preset time period is compared daily with a pre-maintained special event calendar to determine whether the period falls within the occurrence time of any known special event. This determination uses Boolean logic for binary decision-making, and the result determines whether subsequent deviation correction for special events is needed. This ensures that additional correction procedures are only triggered when the period actually covers a special event, avoiding unnecessary adjustments to regular periods without special events.

[0071] Furthermore, when a special event is included in the judgment period, the actual number of complaints related to that special event in the same past time interval is retrieved from the historical complaint database and compared with a preset benchmark number of complaints. The difference between the two is calculated as a deviation value. This deviation value quantifies the additional pull or inhibition effect of the special event on the number of complaints, reflects the incremental contribution of the special event relative to the normal level, and provides data support for the final correction.

[0072] Furthermore, the deviation value is directly added to the revised total number of complaints to complete the final expected total number of complaints output. The essence of this operation is to further superimpose abnormal fluctuations caused by special events on top of the seasonal and holiday fluctuations, so that the final forecast value simultaneously covers both regular cyclical patterns and the impact of occasional events, resulting in a complete expected total number of complaints that takes into account both normal and abnormal conditions.

[0073] In this embodiment of the invention, the total number of complaints in the first set of historical customers and the total number of complaints in the second set of new customers are aggregated to achieve hierarchical fusion prediction of the entire customer group. This avoids the errors caused by using a uniform model due to different customer life cycle stages, greatly improves the throughput and response speed of the large-scale complaint prediction system, and improves the accuracy of business complaint volume analysis.

[0074] S7. Determine the target complaint volume for the target business based on the expected total complaint volume, and visualize the expected total complaint volume and the target complaint volume.

[0075] In this embodiment of the invention, the target complaint volume refers to a complaint control target value determined by a rule engine, optimization algorithm, or manual setting based on the expected total complaint volume and the actual operational constraints of the target business (such as customer service capacity, complaint handling resource limits, performance evaluation indicators, or industry regulatory thresholds), serving as a management or evaluation reference. This value can be an upper limit threshold, a lower limit benchmark, or an expected range. The visualization refers to the technical process of comparing and presenting the expected total complaint volume and the target complaint volume in a graphical interface. Common forms include dual-bar comparison charts, dashboard indicator charts, line trend overlay charts, traffic light warning signs, or difference percentage cards, supporting real-time refresh, interactive drill-down, and dynamic alarm functions.

[0076] In this embodiment of the invention, determining the target complaint volume for the target business based on the expected total complaint volume includes: Obtain the business identifier of the target business, and query the historical average percentage of the target business in the preset total historical complaints based on the business identifier from the preset business complaint percentage mapping table; Multiply the expected total number of complaints by the historical average percentage to obtain the initial allocated number of complaints for the target business; Obtain the business volume change trend of the target business, and determine the business volume correction factor based on the business volume change trend; The target complaint volume for the target business is obtained by multiplying the initial allocated complaint volume by the business volume correction factor.

[0077] In this embodiment of the invention, the business identifier of the target business is obtained and used as an index key to perform a precise lookup in a pre-maintained business complaint percentage mapping table. This mapping table records the average share of each business line in the total historical complaints. The historical average percentage corresponding to the target business is located through the business identifier. This percentage reflects the proportion of the business line's complaint contribution relative to the overall business over a long period of operation, and serves as the basis for allocating the total expected complaint volume down to specific business lines.

[0078] The initial allocation of complaints for the target business is calculated by multiplying the expected total number of complaints by the historical average percentage. Essentially, this operation distributes the total expected number of complaints proportionally across all business lines based on historical experience. The initial allocation represents the complaint distribution for that business line based on historical patterns, without considering current changes in business volume.

[0079] Specifically, data on the current business volume trend of the target business is obtained. This trend reflects the direction and magnitude of recent growth or contraction in the business volume relative to the historical baseline. A business volume correction factor is calculated based on this trend data. This factor quantifies the transmission effect of business volume changes on complaint volume. When business volume increases, the correction factor is greater than the baseline value; when business volume decreases, the correction factor is less than the baseline value, thus incorporating the dynamic changes in business scale into the adjustment framework for complaint volume forecasting.

[0080] Furthermore, the initial allocated complaint volume is multiplied by the business volume adjustment factor to obtain the target complaint volume for the target business. This operation, based on the historical proportion allocation, further scales and adjusts according to the actual changes in the current business volume, so that the final result both inherits the historical allocation weights and reflects the true state of the current business scale. The target complaint volume is the final predicted complaint value allocated to this business line, which can be directly used for subsequent complaint resource allocation and service strategy formulation.

[0081] In this embodiment of the invention, the expected total number of complaints is used as a global total data object, and the target number of complaints is used as a business line-level allocation data object. The two are bound together as a parent-child relationship according to business hierarchy. Simultaneously, a preset time period is used as the horizontal axis dimension, the complaint volume value as the vertical axis dimension, and the business line identifier as the classification dimension, constructing a multi-dimensional data structure that includes total hierarchy, time series, and business classification. This structure provides clear data input and mapping rules for subsequent visualization rendering, ensuring that every visual element in the chart accurately corresponds to the actual business data.

[0082] Specifically, the system automatically matches the optimal visualization chart format based on data characteristics. The expected total number of complaints, due to its time-series attribute, is displayed using a line chart or area chart to show its trend over time. The target number of complaints, due to its multi-business-line comparison attribute, is displayed using a bar chart or grouped bar chart to show the differences in complaint volume between different business lines. The data binding results are converted into graphical drawing instructions recognizable by the rendering engine, including coordinate axis ranges, data point positions, color mapping rules, and legend annotations. This generates a complete drawing blueprint for the final graphical output, allowing users to intuitively perceive both the overall complaint scale and the distribution across different business lines.

[0083] In this embodiment of the invention, the target complaint volume is dynamically determined based on the expected total complaint volume, realizing a closed-loop control from passive prediction to proactive management of thresholds. It can adaptively adjust business KPIs based on real-time data, avoiding the failure of fixed thresholds during operational fluctuations. The expected total complaint volume and the target complaint volume are visualized and compared, and components such as line charts can be used to support interactive drill-down and real-time alarms, helping operations personnel to quickly locate anomalies. At the same time, the visualized difference display facilitates root cause tracing by the system, significantly improving decision-making efficiency and system interpretability.

[0084] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0085] like Figure 3 The diagram shown is a functional block diagram of a business complaint data analysis device provided in an embodiment of the present invention.

[0086] In this embodiment of the disclosure, a business complaint data analysis device is provided, which corresponds one-to-one with the business complaint data analysis method described in the above embodiments. For example... Figure 3 As shown, this business complaint data analysis device 100 can be installed in an electronic device. According to its functions, the business complaint data analysis device 100 includes a structured processing module 101, a customer set segmentation module 102, a first complaint volume calculation module 103, a complaint factor extraction module 104, a second complaint volume calculation module 105, an expected complaint volume analysis module 106, and a visualization display module 107. Detailed descriptions of each functional module are as follows: The structured processing module 101 is used to obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset. The customer set segmentation module 102 is used to identify the identity identifiers of the complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaining customer identity identifiers and a preset time reference point as the boundary. The first complaint volume calculation module 103 is used to construct the re-complaint probability function for each historical complaint customer in the historical complaint customer set, and calculate the first total number of complaints from the historical complaint customers in a future preset time period based on the re-complaint probability function. The complaint factor extraction module 104 is used to extract the complaint baseline quantity of each new complainant in the new complaint customer set, and to collect various complaint event influencing factors of the new complaint customer set within the preset time period. The second complaint volume calculation module 105 is used to correct the complaint baseline volume based on various complaint event influencing factors to obtain the second total number of complaints of the newly added complaint customer set within the preset time period. The expected complaint volume analysis module 106 is used to summarize the first total complaint volume and the second total complaint volume to generate the expected total complaint volume within the preset time period. The visualization module 107 is used to determine the target complaint volume of the target business based on the expected total complaint volume, and to visualize the expected total complaint volume and the target complaint volume.

[0087] In one embodiment, when the customer set segmentation module 102 performs the task of identifying the identity of complaining customers in the target complaint dataset, it is used to: Obtain each complaint record in the target complaint dataset, and extract at least one preset identity association field from the complaint record; The identity association field is formatted and standardized to obtain the target identity association field; Each of the target identity association fields is matched with the corresponding field in the preset customer profile database to obtain the confidence value of each target identity association field that is successfully matched; The target identity association field corresponding to the matching result whose confidence value is greater than or equal to the preset confidence threshold is determined as the complainant customer identity identifier in the complaint record.

[0088] In one embodiment, the first complaint volume calculation module 103, when executing the execution of constructing the re-probability function for each historical complaint customer in the historical complaint customer set, is used to: Extract the complaint time series of each historical complaint customer from each complaint record in the historical complaint customer set, and calculate the time interval between two adjacent complaint records based on the complaint time series to obtain the complaint interval sequence; Using the preset maximum likelihood estimation method, the parameters of at least two preset candidate probability distribution models are fitted according to the complaint interval sequence to obtain the fitted probability distribution model. Calculate the goodness-of-fit index between each of the fitted probability distribution models and the complaint interval sequence; Based on the goodness-of-fit index, the optimal probability distribution model is selected from the fitted probability distribution models; The cumulative distribution function corresponding to the optimal probability distribution model is determined as the re-suit probability function for the historical complaint customers.

[0089] In one embodiment, when the first complaint volume calculation module 103 calculates the total number of first complaints from the historical complaint customers within a future preset time period based on the re-complaint probability function, it is used to: Obtain the duration of a future preset time period, and construct a prediction time window based on the duration of the period; The prediction time window is divided into multiple consecutive and non-overlapping unit time intervals; The probability of a historical complaint customer filing at least one complaint within the unit time interval is calculated based on the re-complaint probability function. The interval complaint probabilities corresponding to all the unit time intervals are summed, and the sum is rounded up to obtain the first total number of complaints from the historical complaint customers in the future preset time period.

[0090] In one embodiment, when the second complaint volume calculation module 105 performs a correction calculation on the complaint baseline volume based on various complaint event influencing factors to obtain the second total number of complaints for the newly added customer set within the preset time period, it is used to: Obtain the time series of the influence intensity of each of the complaint event's influence factors over time within the preset time period, and align the time series of influence intensity on the time axis to generate an influence factor matrix; Match the set of sensitivity coefficients of the newly added complaining customer to each of the factors affecting the complaint event from the preset factor sensitivity configuration table; Each sensitivity coefficient in the sensitivity coefficient set is multiplied point by point with the corresponding influence intensity time series in the influence factor matrix to obtain the total correction coefficient for the newly added complaining customers; Multiplying the baseline number of complaints by the total correction factor yields the second total number of complaints from the newly added customer set within the preset time period.

[0091] In one embodiment, when the expected complaint volume analysis module 106 performs the process of summarizing the first total complaint volume and the second total complaint volume to generate the expected total complaint volume within the preset time period, it is used to: The initial total number of complaints is obtained by summing the first total number of complaints and the second total number of complaints. Obtain seasonal and holiday characteristic information contained within the preset time period, and make a first correction to the initial total number of complaints based on the seasonal and holiday characteristic information to obtain the corrected total number of complaints; Determine whether a preset special event is included within the preset time period; If included, the historical complaint volume of the special event during the same historical time period is extracted from the preset historical complaint database, and the deviation value between the historical complaint volume and the preset benchmark complaint volume is calculated. The deviation value is added to the corrected total number of complaints to obtain the expected total number of complaints within the preset time period.

[0092] In one embodiment, when the visualization module 107 determines the target complaint volume for the target business based on the expected total complaint volume, it is used to: Obtain the business identifier of the target business, and query the historical average percentage of the target business in the preset total historical complaints based on the business identifier from the preset business complaint percentage mapping table; Multiply the expected total number of complaints by the historical average percentage to obtain the initial allocated number of complaints for the target business; Obtain the business volume change trend of the target business, and determine the business volume correction factor based on the business volume change trend; The target complaint volume for the target business is obtained by multiplying the initial allocated complaint volume by the business volume correction factor.

[0093] In this invention, the specific limitations of the business complaint data analysis device can be found in the above-described limitations of the business complaint data analysis method, and will not be repeated here. Each module in the aforementioned business complaint data analysis device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0094] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the business complaint data analysis method on the server side.

[0095] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the business complaint data analysis method on the client side.

[0096] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset; Identify the identity of complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaint customer identity and a preset time reference point. Construct a re-complaint probability function for each historical complainant in the historical complainant set, and calculate the first total number of complaints from the historical complainant in a future preset time period based on the re-complaint probability function; Extract the baseline complaint volume for each new customer in the newly added complaint customer set, and collect various complaint event influencing factors for the newly added complaint customer set within the preset time period; The baseline complaint volume is adjusted and calculated based on various factors affecting the complaint events to obtain the second total complaint volume of the newly added customer set within the preset time period. The total number of the first complaint and the total number of the second complaint are combined to generate the expected total number of complaints within the preset time period; The target complaint volume for the target business is determined based on the expected total complaint volume, and the expected total complaint volume and the target complaint volume are visualized.

[0097] In the several embodiments provided by this invention, it should be understood that the disclosed devices and apparatuses can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0098] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0099] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0100] In some embodiments of this example, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method described in the above embodiments.

[0101] The readable storage medium of the present invention stores a computer program, which, when executed by a processor of an electronic device, can perform the following: Obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset; Identify the identity of complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaint customer identity and a preset time reference point. Construct a re-complaint probability function for each historical complainant in the historical complainant set, and calculate the first total number of complaints from the historical complainant in a future preset time period based on the re-complaint probability function; Extract the baseline complaint volume for each new customer in the newly added complaint customer set, and collect various complaint event influencing factors for the newly added complaint customer set within the preset time period; The baseline complaint volume is adjusted and calculated based on various factors affecting the complaint events to obtain the second total complaint volume of the newly added customer set within the preset time period. The total number of the first complaint and the total number of the second complaint are combined to generate the expected total number of complaints within the preset time period; The target complaint volume for the target business is determined based on the expected total complaint volume, and the expected total complaint volume and the target complaint volume are visualized.

[0102] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0103] Computer-readable storage media may also store at least one computer-executable program / instruction, such as computer-readable instructions. Computer-readable storage media include, but are not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Computer-readable storage media may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions stored on the computer-readable storage medium, the various methods described above can be performed.

[0104] In addition, the computer device may include (but is not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (e.g., keyboard, mouse, speakers, etc.).

[0105] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product / computer program product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0106] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Furthermore, any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory.

[0107] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0108] In the embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0109] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

[0110] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

Claims

1. A method for analyzing business complaint data, characterized in that, The method includes: Obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset; Identify the identity of complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaint customer identity and a preset time reference point. Construct a re-complaint probability function for each historical complainant in the historical complainant set, and calculate the first total number of complaints from the historical complainant in a future preset time period based on the re-complaint probability function; Extract the baseline complaint volume for each new customer in the newly added complaint customer set, and collect various complaint event influencing factors for the newly added complaint customer set within the preset time period; The baseline complaint volume is adjusted and calculated based on various factors affecting the complaint events to obtain the second total complaint volume of the newly added customer set within the preset time period. The total number of the first complaint and the total number of the second complaint are combined to generate the expected total number of complaints within the preset time period; The target complaint volume for the target business is determined based on the expected total complaint volume, and the expected total complaint volume and the target complaint volume are visualized.

2. The business complaint data analysis method as described in claim 1, characterized in that, The construction of the re-suit probability function for each historical complaint customer in the historical complaint customer set includes: Extract the complaint time series of each historical complaint customer from each complaint record in the historical complaint customer set, and calculate the time interval between two adjacent complaint records based on the complaint time series to obtain the complaint interval sequence; Using the preset maximum likelihood estimation method, the parameters of at least two preset candidate probability distribution models are fitted according to the complaint interval sequence to obtain the fitted probability distribution model. Calculate the goodness-of-fit index between each of the fitted probability distribution models and the complaint interval sequence; Based on the goodness-of-fit index, the optimal probability distribution model is selected from the fitted probability distribution models; The cumulative distribution function corresponding to the optimal probability distribution model is determined as the re-suit probability function for the historical complaint customers.

3. The business complaint data analysis method as described in claim 1, characterized in that, The step of calculating the total number of complaints from historical complainants within a future preset time period based on the re-complaint probability function includes: Obtain the duration of a future preset time period, and construct a prediction time window based on the duration of the period; The prediction time window is divided into multiple consecutive and non-overlapping unit time intervals; The probability of a historical complaint customer filing at least one complaint within the unit time interval is calculated based on the re-complaint probability function. The interval complaint probabilities corresponding to all the unit time intervals are summed, and the sum is rounded up to obtain the first total number of complaints from the historical complaint customers in the future preset time period.

4. The business complaint data analysis method as described in claim 1, characterized in that, The step of adjusting the baseline complaint volume based on various influencing factors of the complaint events to obtain the second total number of complaints for the newly added customer set within the preset time period includes: Obtain the time series of the influence intensity of each of the complaint event's influence factors over time within the preset time period, and align the time series of influence intensity on the time axis to generate an influence factor matrix; Match the set of sensitivity coefficients of the newly added complaining customer to each of the factors affecting the complaint event from the preset factor sensitivity configuration table; Each sensitivity coefficient in the sensitivity coefficient set is multiplied point by point with the corresponding influence intensity time series in the influence factor matrix to obtain the total correction coefficient for the newly added complaining customers; Multiplying the baseline number of complaints by the total correction factor yields the second total number of complaints from the newly added customer set within the preset time period.

5. The business complaint data analysis method as described in claim 1, characterized in that, The step of summing the first total number of complaints and the second total number of complaints to generate the expected total number of complaints within the preset time period includes: The initial total number of complaints is obtained by summing the first total number of complaints and the second total number of complaints. Obtain seasonal and holiday characteristic information contained within the preset time period, and make a first correction to the initial total number of complaints based on the seasonal and holiday characteristic information to obtain the corrected total number of complaints; Determine whether a preset special event is included within the preset time period; If included, the historical complaint volume of the special event during the same historical time period is extracted from the preset historical complaint database, and the deviation value between the historical complaint volume and the preset benchmark complaint volume is calculated. The deviation value is added to the corrected total number of complaints to obtain the expected total number of complaints within the preset time period.

6. The business complaint data analysis method as described in claim 1, characterized in that, Determining the target complaint volume for the target business based on the expected total complaint volume includes: Obtain the business identifier of the target business, and query the historical average percentage of the target business in the preset total historical complaints based on the business identifier from the preset business complaint percentage mapping table; Multiply the expected total number of complaints by the historical average percentage to obtain the initial allocated number of complaints for the target business; Obtain the business volume change trend of the target business, and determine the business volume correction factor based on the business volume change trend; The target complaint volume for the target business is obtained by multiplying the initial allocated complaint volume by the business volume correction factor.

7. The business complaint data analysis method as described in claim 1, characterized in that, The identification of the complainant's identity in the target complaint dataset includes: Obtain each complaint record in the target complaint dataset, and extract at least one preset identity association field from the complaint record; The identity association field is formatted and standardized to obtain the target identity association field; Each of the target identity association fields is matched with the corresponding field in the preset customer profile database to obtain the confidence value of each target identity association field that is successfully matched; The target identity association field corresponding to the matching result whose confidence value is greater than or equal to the preset confidence threshold is determined as the complainant customer identity identifier in the complaint record.

8. A business complaint data analysis device, characterized in that, The device includes: The structured processing module is used to obtain the historical complaint dataset of the target business, perform structured processing on the historical complaint dataset, and generate the target complaint dataset. The customer set segmentation module is used to identify the identity of the complaining customers in the target complaint dataset, and divide the target complaint dataset into a historical complaint customer set and a newly added complaint customer set based on the complaining customer identity and a preset time reference point. The first complaint volume calculation module is used to construct the re-complaint probability function for each historical complaint customer in the historical complaint customer set, and calculate the first total number of complaints from the historical complaint customers in a future preset time period based on the re-complaint probability function. The complaint factor extraction module is used to extract the complaint baseline quantity for each new customer in the new complaint customer set, and to collect various complaint event influencing factors of the new complaint customer set within the preset time period. The second complaint volume calculation module is used to correct the complaint baseline volume based on various complaint event influencing factors to obtain the second total number of complaints for the newly added complaint customer set within the preset time period. The expected complaint volume analysis module is used to summarize the first total complaint volume and the second total complaint volume to generate the expected total complaint volume within the preset time period; The visualization module is used to determine the target complaint volume for the target business based on the expected total complaint volume, and to visualize the expected total complaint volume and the target complaint volume.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the business complaint data analysis method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the business complaint data analysis method as described in any one of claims 1 to 7.

Images