Search

CN-122022824-A - Customer satisfaction feedback collection method and system based on community collaborative filtering

CN122022824ACN 122022824 ACN122022824 ACN 122022824ACN-122022824-A

Abstract

The invention relates to the technical field of business intelligence and data mining, in particular to a customer satisfaction feedback collection method and system based on community collaborative filtering; the method comprises the steps of constructing a customer service experience track through an analysis service flow, dynamically dividing service experience communities according to similarities of behaviors, time and abnormal states, generating a stage satisfaction baseline and a reliability index based on feedback data existing in communities, intelligently screening high-representative customers through a collaborative filtering algorithm by combining the communities baseline, the stage reliability and the customer behavior similarity, determining optimal acquisition opportunities to achieve accurate touch, predicting satisfaction of the non-fed-back customers and comparing the non-fed-back customers with the baseline to determine whether to complement the acquired customers, and updating the communities baseline and an analysis model after verification and enhancement of the collected feedback, and finally outputting a structural satisfaction conclusion through steady-state analysis, sensitivity reconstruction and consistency constraint. The invention effectively improves the feedback collection efficiency, the data quality and the decision support capability.

Inventors

  • LI MENG

Assignees

  • 北京易享信息技术有限公司

Dates

Publication Date
20260512
Application Date
20260203

Claims (10)

  1. 1. The customer satisfaction feedback collection method based on community collaborative filtering is characterized by comprising the following specific implementation steps: s1, constructing a customer service experience track by analyzing a service flow, dynamically dividing service experience communities according to behavioral response, time experience and similarity of abnormal states, and generating community service state labels; S2, based on feedback data existing in communities, carrying out acquisition pretreatment, stage statistical analysis, baseline construction and dynamic update to generate community satisfaction perception baselines comprising stage satisfaction baselines and stage credibility indexes; S3, combining community satisfaction perception baselines, stage credibility and customer behavior tracks, screening high-representative customers and key stages by calculating customer contribution degree and stage trigger factors, predicting non-feedback customer satisfaction degree by utilizing collaborative filtering, and finally generating feedback acquisition object priority and trigger strategies; S4, according to the feedback acquisition priority and the triggering strategy, personalized feedback request sending, real-time feedback receiving and checking are carried out, the reliability data with the missing reliability lower than the set threshold value is synergistically enhanced and complemented with the base line by utilizing the predicted value, and the community satisfaction perception base line is updated; S5, carrying out steady-state analysis, stage sensitivity reconstruction and community consistency constraint processing on the enhanced community satisfaction data, outputting a structured comprehensive satisfaction conclusion, and using the conclusion to guide subsequent feedback collection and optimization.
  2. 2. The method for collecting customer satisfaction feedback based on collaborative filtering according to claim 1, wherein in step S1, constructing a customer service experience track specifically includes: according to a service flow structure predefined by a service system, the service process is disassembled into a plurality of service stages with definite service meanings; Continuously collecting operation behaviors, interaction responses, residence time and abnormal state information of a client in each service stage; And organizing and merging the discrete behavior information according to the sequence of the service stages to form a service experience track of the client.
  3. 3. The method for collecting customer satisfaction feedback based on collaborative filtering according to claim 2, wherein in step S1, dynamically dividing service experience communities specifically includes: On the premise of ensuring that service experience tracks of different clients are in the same service stage, calculating similarity between clients on behavior response characteristics, consistency in residence time and response time length and consistency in abnormal state occurrence conditions; Introducing a preset importance weight of a service stage, and carrying out weighted synthesis on similarity calculation results of the three dimensions to obtain overall similarity reflecting service experience consistency among clients; Based on the overall similarity calculation result, under the constraint of the dynamically adjusted consistency threshold, clients with similar experience heights are gradually aggregated to form a service experience community.
  4. 4. The method for collecting feedback of customer satisfaction based on collaborative filtering according to claim 3, wherein in step S2, the generating process of the community satisfaction awareness base line specifically includes: identifying and collecting customer data which are fed back in communities, and unifying and standardizing original feedback values of different sources and formats into values in a preset range; Carrying out weighted interpolation and adjustment treatment on feedback values with defects or anomalies by combining stage weights and anomaly marks; based on the processed feedback data, calculating the average satisfaction degree and the satisfaction degree standard deviation of communities in each service stage in a statistics mode; analyzing the change trend of satisfaction degree of each stage with time or flow promotion, and generating a stage trend index; Combining the stage weights to construct a stage satisfaction baseline vector reflecting the overall experience mode of the community; Meanwhile, a credibility index is calculated for each stage, and the index comprehensively reflects the coverage sufficiency and internal consistency of the feedback data of the stage.
  5. 5. The method for collecting customer satisfaction feedback based on collaborative community filtering according to claim 4, wherein in step S3, screening highly representative customers specifically comprises: For each client in the community, calculating the comprehensive similarity between the client and the baseline value of each stage in the community satisfaction perception baseline; Combining the credibility indexes of each stage to generate a customer contribution index for measuring the representative degree of the customer to the overall satisfaction of the community; And sorting all clients in the community according to the client contribution index, and screening clients with contribution index higher than a set threshold value to form a high-representative client set as an object for feedback collection preferentially.
  6. 6. The method for collecting feedback of customer satisfaction based on collaborative filtering according to claim 5, wherein in step S3, the determining machine of feedback collection specifically comprises: Aiming at each service stage, calculating to obtain a stage trigger factor according to the stability of a stage satisfaction baseline, the significance degree of a stage satisfaction change trend and the level of a stage reliability index; when the trigger factor of a certain service stage exceeds a preset trigger threshold, judging the stage as a key stage requiring feedback acquisition at present, and incorporating the key stage into a trigger stage set; The feedback acquisition request is initiated to the client only within the service phase comprised by the set of trigger phases.
  7. 7. The method for collecting feedback of customer satisfaction based on collaborative filtering according to claim 6, wherein in step S3, predicting the unfeeded customer satisfaction using collaborative filtering specifically comprises: For clients which do not provide feedback in communities, predicting satisfaction scores of the clients in a key service stage according to the behavior similarity of the clients and the fed back clients on the service experience track by using a collaborative filtering algorithm based on the neighborhood; comparing the predicted satisfaction score with a community satisfaction baseline of a corresponding service stage, and calculating the difference between the two; if the difference exceeds a preset difference threshold, determining that the experience of the non-feedback client may deviate significantly from the community baseline, and triggering an active feedback acquisition request for the client.
  8. 8. The method for collecting customer satisfaction feedback based on collaborative community filtering according to claim 7, wherein in step S4, collaborative enhancement and completion of feedback data specifically comprises: For the lack of satisfaction data caused by the fact that the customer does not feed back or the stage reliability is lower than a set threshold value, the predicted satisfaction value obtained in the step S3 and the corresponding stage satisfaction baseline value are used for data filling in a weighted fusion mode; Calculating a confidence index for filling the generated satisfaction data to represent the reliability of the completion data; and integrating the actually collected effective feedback data with the filled data to form an enhanced community satisfaction matrix.
  9. 9. The method for collecting customer satisfaction feedback based on collaborative community filtering according to claim 8, wherein outputting a structured integrated satisfaction conclusion in step S5 comprises: Extracting data from the enhanced community satisfaction matrix, performing steady-state analysis on a plurality of customer feedback of each service stage, and removing extreme value influence to obtain a stage steady-state satisfaction value; Analyzing historical fluctuation trend and discrete degree of satisfaction of each stage, and dynamically determining weight of each stage in overall satisfaction evaluation through a sensitivity reconstruction mechanism; And finally outputting a structured conclusion packet which at least comprises the comprehensive satisfaction value of the community, a key service stage list with obvious influence on the whole satisfaction and a state label reflecting the community internal experience consistency.
  10. 10. A customer satisfaction feedback collection system based on community collaborative filtering, for executing the customer satisfaction feedback collection method based on community collaborative filtering according to any one of claims 1 to 9, characterized in that a cloud edge collaborative architecture is adopted, comprising: the community self-adaptive construction module is deployed at the edge node and is used for collecting client behavior data in real time, constructing a local service experience track and forming a local community prototype; the community baseline and stage credibility generation module is deployed at the cloud end and is used for fusing polygonal node data to generate and maintain a global community satisfaction perception baseline model; The feedback acquisition object and trigger strategy generation module is deployed at the edge node and is used for generating a local feedback acquisition priority and a trigger instruction according to the baseline model issued by the cloud and the local real-time state; the client feedback collection and data enhancement module is deployed at the edge node and is used for executing local feedback request sending and receiving verification and carrying out local enhancement and completion on missing data; And the community satisfaction comprehensive analysis and result output module is deployed at the cloud and used for integrating global enhancement data, carrying out steady-state analysis, sensitivity reconstruction and consistency verification and outputting comprehensive satisfaction conclusion and optimization strategies.

Description

Customer satisfaction feedback collection method and system based on community collaborative filtering Technical Field The invention relates to the technical field of business intelligence and data mining, in particular to a customer satisfaction feedback collection method and system based on community collaborative filtering. Background With the continuous deepening of customer experience management and the popularization of data-driven decision modes, the requirements of enterprises on the instantaneity, representativeness and operability of customer satisfaction feedback are increasingly improved, and under the technical background of service flow digitization and traceability of customer behaviors, how to efficiently integrate multi-dimensional customer interaction data and extract high-quality experience insight from the multi-dimensional customer interaction data becomes an important subject. The invention discloses a customer service intelligent interaction method, system and medium based on AI, which comprises collecting attribute information and behavior track information of a target customer, processing to obtain customer information portraits, extracting personalized demand information of the target customer according to the customer information portraits, executing intelligent interaction service based on the personalized demand information, monitoring the interaction service process in real time, collecting service state parameters, judging whether manual customer service switching is required to be executed or not according to the service state parameters, monitoring intelligent interaction service within a preset time period, extracting service index data, carrying out quantitative evaluation processing according to the service index data to obtain service reliability coefficients, judging the reliability of the intelligent interaction service and adopting corresponding optimization measures, thereby realizing the technology of intelligent customer service interaction based on AI. The collaborative filtering technology is used as an effective means for mining the group preference rule, has proved value in the fields of recommendation and the like, provides technical heuristic for introducing group similarity analysis in customer satisfaction management, and simultaneously promotes the development of a feedback collection method in a more intelligent, more accurate and more self-adaptive direction together with the evolution of related technologies such as dynamic community construction, time sequence data analysis and personalized trigger mechanism. Disclosure of Invention The invention aims at solving the problems in the background technology and provides a customer satisfaction feedback collection method and system based on community collaborative filtering. The technical scheme of the invention is that the customer satisfaction feedback collection method based on community collaborative filtering comprises the following specific implementation steps: s1, constructing a customer service experience track by analyzing a service flow, dynamically dividing service experience communities according to behavioral response, time experience and similarity of abnormal states, and generating community service state labels; S2, based on feedback data existing in communities, carrying out acquisition pretreatment, stage statistical analysis, baseline construction and dynamic update to generate community satisfaction perception baselines comprising stage satisfaction baselines and stage credibility indexes; S3, combining community satisfaction perception baselines, stage credibility and customer behavior tracks, screening high-representative customers and key stages by calculating customer contribution degree and stage trigger factors, predicting non-feedback customer satisfaction degree by utilizing collaborative filtering, and finally generating feedback acquisition object priority and trigger strategies; S4, according to the feedback acquisition priority and the triggering strategy, personalized feedback request sending, real-time feedback receiving and checking are carried out, the reliability data with the missing reliability lower than the set threshold value is synergistically enhanced and complemented with the base line by utilizing the predicted value, and the community satisfaction perception base line is updated; S5, carrying out steady-state analysis, stage sensitivity reconstruction and community consistency constraint processing on the enhanced community satisfaction data, outputting a structured comprehensive satisfaction conclusion, and using the conclusion to guide subsequent feedback collection and optimization. Preferably, in step S1, constructing a customer service experience track specifically includes: according to a service flow structure predefined by a service system, the service process is disassembled into a plurality of service stages with definite service meanings; Continuous