SCRM-based data analysis model

Abstract

The invention relates to the technical field of data analysis models, and particularly discloses an SCRM-based data analysis model, which is used for acquiring customer data from social media, a customer relationship management system and customer feedback; cleaning and formatting the collected customer data, extracting feature data related to after-sales service, and performing data fusion; based on an output result of the data fusion model, establishing an after-sales service demand prediction model, and predicting a future after-sales service demand change trend of the customer; the distribution of after-sales service resources is optimized according to the after-sales service demand calculated by the after-sales service demand prediction model; and a customer demand trend and a resource allocation scheme are displayed in a graphical mode. According to the method, the multi-source data is integrated, the data island is broken, the resource utilization rate is maximized and the customer waiting time is minimized by optimizing the objective function, and the user satisfaction is improved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis models, and more specifically, to a data analysis model based on SCRM. Background Art

[0002] With the widespread application of customer relationship management (CRM) technology, enterprises have strong capabilities in obtaining basic customer information and historical behavior data, but the traditional CRM model has the problem of insufficient response to real-time dynamic customer needs. With the popularity of social media, customers are increasingly using social networks to feedback product experience, express after-sales needs or raise questions. The real-time and diversity of these data make them an important supplement to customer relationship management. However, existing CRM systems often cannot effectively integrate these unstructured social data, lack the ability to deeply analyze customer behavior and potential needs, and are particularly insufficient in predicting the changing trend of customer after-sales service needs. At the same time, enterprises still mainly respond passively in the allocation of after-sales service resources, and are unable to deploy resources in advance, resulting in long customer waiting time, reduced satisfaction, and possible customer churn. Therefore, how to use data analysis to deeply mine customer behavior and social data based on the concept of social customer relationship management (SCRM), predict the changing trend of customer after-sales service needs in real time, and provide solutions to optimize resource allocation has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a data analysis model based on SCRM, which collects customer data from social media, CRM systems and customer feedback, establishes an after-sales service demand prediction model to predict the future trend of customer demand changes, and then optimizes the allocation of after-sales service resources based on the prediction results. At the same time, it displays customer demand trends and resource allocation plans in a graphical manner, integrates multi-source data, breaks data silos, and maximizes resource utilization and minimizes customer waiting time by optimizing the objective function, thereby improving user satisfaction and solving the problems raised in the above-mentioned background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A data analysis model based on SCRM includes the following steps:

[0006] Step S1, collecting customer data from social media, CRM system and customer feedback;

[0007] Step S2, combining the customer data after cleaning and formatting operations and extracting the after-sales service related feature data to obtain the score D after data fusion, the contribution value of social media customer data to the fusion score is Among them, Si is the i-th piece of social media data, T is the extracted after-sales service feature data, n is the number of social media data currently being processed, N S is the total amount of social media data, t is the current time point, and t i is the timestamp of the ith social media data, α is the weight coefficient of social media data in the overall fusion, γ is the smoothing factor of social media data, and controls the total amount of social media data N S The adjustment coefficient for the score after data fusion, β is the first time decay rate parameter;

[0008] Step S3, based on the score after the data fusion, establish an after-sales service demand prediction model to predict the customer's future after-sales service demand. The formula of the after-sales service demand prediction model is:

[0009]

[0010] Where: P(s) is the customer's after-sales service demand at the future time point s; D is the score after data fusion calculated by the data fusion model; s is the predicted time point; s0 is the current time; φ is the time decay rate; r is the total number of complaints submitted by customers; s k The time when the customer submits the complaint; ΔD k is the demand fluctuation, ΔD k satisfy in, is the time point s k The score after data fusion is is the time point s k The score after data fusion at the previous time point;

[0011] Step S4, optimizing the allocation of after-sales service resources according to the demand for after-sales service;

[0012] Step S5, displaying customer demand trends and resource allocation plans in a graphical manner.

[0013] As a further solution of the present invention, the contribution value of CRM system customer data to the fusion score is Among them, C j is the jth CRM system data, m is the number of data in the CRM system; C avg is the average value of all relevant data feature values ​​in the CRM system, t j is the timestamp of the CRM system data, δ is the weight coefficient of the data in the CRM system in the overall fusion, and λ is the second time decay rate parameter; the contribution of customer feedback data to the fusion score is Among them, p is the number of customer feedback data items currently used, F kis the kth customer feedback data, N F is the total amount of customer feedback data, ζ is the smoothing factor of customer feedback data, and controls the total amount of customer feedback data N F The adjustment coefficient that affects the score after data fusion, η is the weight coefficient of customer feedback data in the overall fusion.

[0014] As a further solution of the present invention, customer data is collected from social media, CRM systems and customer feedback, including the following specific contents: First, customer data from social media platforms are integrated, and the customer data includes customer comments, likes, shares, replies and other interactive behaviors, especially feedback information related to product failures. For example, when customers publicly discuss or raise questions about product usage issues on social platforms, text data can reflect the customer's real needs and emotional state for the product or service. In order to capture the text data, data is connected to mainstream social media platforms through an application program interface to dynamically monitor customer activities in different time periods. At the same time, natural language processing is used to identify the category, severity and emotional tendency of the fault problem through keyword extraction and sentiment analysis in customer comments.

[0015] Secondly, connect with the CRM system to obtain the customer's historical behavior data and structured information, including the customer's purchase record, transaction frequency, detailed information of the purchased products, and interaction records with the customer service department.

[0016] In addition, it also covers the data sources of customer feedback channels, especially the feedback information actively provided by customers through e-mail, online customer service, telephone or questionnaire surveys, etc. The feedback information includes customers' subjective evaluation of product performance and service quality, which can supplement the content not covered by social media and CRM systems. In order to efficiently integrate diverse data, a distributed data collection architecture is adopted, and data from different channels are processed consistently through a unified interface and data format standardization technology to ensure the integrity and accuracy of data input.

[0017] As a further solution of the present invention, the collected customer data is cleaned and formatted, and the characteristic data related to after-sales service is extracted, and then data fusion is performed, including the following specific contents: First, the data collected from social media, CRM system and customer feedback is cleaned, and the cleaning operation includes: deleting invalid data or duplicate data, filling missing values ​​and removing outliers. The data is marked and filtered by a deduplication algorithm. For the processing of missing values ​​in the data, a variety of filling strategies are adopted, and the average value or median is used to fill the numerical data; the nearest neighbor algorithm is used to predict and fill the categorical data.

[0018] After cleaning, the data is formatted to convert data from multiple sources into a standardized format. Natural language processing technology is used to convert unstructured text data into analyzable features.

[0019] Based on text analysis technology, feature data related to after-sales service is extracted, and sentiment analysis algorithms are used to classify customer comments into positive, neutral or negative sentiments. The sentiment classification helps to identify the customer's satisfaction level with the current after-sales service. In addition, word frequency analysis and keyword extraction algorithms are used to identify words in the comments that are closely related to after-sales service, and semantic analysis technology is used to associate them with specific customer demand scenarios.

[0020] As a further solution of the present invention, according to the after-sales service demand, the allocation of after-sales service resources is optimized, including the following specific contents: based on the after-sales service demand prediction model, taking the after-sales service demand prediction as input, by optimizing the objective function, the goal of maximizing resource utilization and minimizing customer waiting time is achieved, and the formula of the objective function is:

[0021]

[0022] Where: M is the total number of after-sales service sites; W i is the average waiting time of customers at the i-th service station; P i is the after-sales service demand of the i-th service station; is the average waiting time of customers at the i-th service station W i The after-sales service demand of the i-th service station P i relationship; R i is the amount of resources allocated to the i-th service site; is the amount of resources R allocated to the i-th service site i The matching degree of after-sales service demand with the i-th service station; μ is the balance factor, which controls the balance between minimizing waiting time and resource matching; at the same time, in order to ensure that the total amount of resources allocated to each service station does not exceed the total amount of enterprise resources R total , to constrain the total amount of resources, the constraint formula is: And R i Satisfy R i ≥0, The objective function comprehensively considers the total amount of resources of the enterprise and the needs of each site. First, it takes the after-sales service demand provided by the demand forecasting model as the core, combines the resource allocation needs of each service site, and dynamically adjusts the allocation strategy. At the same time, it considers the customer waiting time and resource matching deviation, ensuring that the number of allocated resources can meet the site needs and avoid resource waste or shortage, and effectively reducing the waiting time of customers caused by unreasonable resource allocation. In addition, the total amount of resource constraints and non-negative constraints are added to ensure that the optimization plan is feasible under realistic constraints, and to achieve a balance between resource utilization and service efficiency through balancing factors.

[0023] As a further solution of the present invention, customer demand trends and resource allocation plans are displayed in a graphical manner, including the following specific contents: by integrating the output results of the after-sales service demand prediction model and the optimization objective function, a variety of visualization means are used to display information, including line graphs, bar graphs, pie charts, heat maps and interactive dashboards.

[0024] Step Y1: graphically display the customer demand trend, use a time series line chart to show the changing trend of customer demand in different time periods, and combine the forecast results to show the fluctuation range and growth point of future demand, so as to help enterprises predict the service peak period in advance;

[0025] Step Y2: Display resource allocation suggestions in a bar chart or stacked chart, specifying the resource allocation quantity for each service site, and marking the resource utilization rate with color coding to facilitate the discovery of potential problems in resource allocation;

[0026] Step Y3, providing a regional heat map function to display the spatial distribution of customer demand based on geographic location, helping enterprises identify demand hotspots and adjust resource layout;

[0027] Step Y5: In order to improve the user experience, a dynamic interactive function is also provided. Users can view detailed data of specific sites, such as customer waiting time, demand satisfaction rate and remaining resources, by clicking, dragging or filtering operations; and integrate key indicators and trend information into analysis results with graphics and texts to generate customized reports, so that decision makers can quickly obtain a global overview or conduct in-depth analysis of specific issues.

[0028] Through the above steps, enterprises can fully understand the current situation and future trends of customer needs and resource allocation, improve decision-making efficiency, optimize after-sales service processes, and ultimately enhance customer satisfaction and enterprise operational efficiency.

[0029] The technical effects and advantages of a data analysis model based on SCRM in the present invention are as follows: the present invention collects customer data from social media, CRM systems and customer feedback, and uses a variety of technical means to clean, format, extract features and fuse the data, establishes an after-sales service demand prediction model to predict future customer demand change trends, and then optimizes after-sales service resource allocation based on the prediction results, while displaying customer demand trends and resource allocation plans in a graphical manner. It can effectively integrate multi-source data, break data silos, deeply mine customer behavior and social data, predict after-sales service demand change trends, deploy resources in advance, reduce customer waiting time, improve customer satisfaction, and reduce customer churn risks; maximize resource utilization and minimize customer waiting time by optimizing the objective function, and consider a variety of realistic constraints to ensure the feasibility of the plan; graphical display facilitates enterprises to fully understand customer needs and resource allocation, improve decision-making efficiency, optimize after-sales service processes, and enhance enterprise operational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a method flow chart of a data analysis model based on SCRM in the present invention. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] Example 1

[0033] See also Figure 1 As shown in the method flow chart, the embodiment of the present invention provides a data analysis model based on SCRM, which includes the following steps:

[0034] Step S1, collecting customer data from social media, CRM system and customer feedback;

[0035] Step S2, combining the customer data after the cleaning and formatting operations and extracting feature data related to after-sales service to obtain a score after data fusion;

[0036] Step S3, based on the scores after the data fusion, establish an after-sales service demand prediction model to predict the customer's future after-sales service demand;

[0037] Step S4, optimizing the allocation of after-sales service resources according to the demand for after-sales service;

[0038] Step S5, displaying customer demand trends and resource allocation plans in a graphical manner.

[0039] Furthermore, customer data is collected from social media, CRM systems and customer feedback, including the following specific contents: First, customer data from social media platforms is integrated, including customer comments, likes, shares, replies and other interactive behaviors, especially feedback information related to product failures. For example, when customers publicly discuss or raise questions about product usage issues on social platforms, text data can reflect the customer's real needs and emotional state for products or services. In order to capture the text data, data is connected to mainstream social media platforms through application programming interfaces to dynamically monitor customer activities in different time periods. At the same time, natural language processing is used to identify the category, severity and emotional tendency of the fault problem through keyword extraction and sentiment analysis in customer comments.

[0040] Secondly, connect with the CRM system to obtain the customer's historical behavior data and structured information, including the customer's purchase history, transaction frequency, details of the purchased products, and interaction records with the customer service department. For example, the customer's repair request, consultation record, complaint history, and service rating information are all important inputs.

[0041] In addition, it also covers the data sources of customer feedback channels, especially the feedback information actively provided by customers through e-mail, online customer service, telephone or questionnaire surveys, etc. The feedback information includes customers' subjective evaluation of product performance and service quality, which can supplement the content not covered by social media and CRM systems. In order to efficiently integrate diverse data, a distributed data collection architecture is adopted, and data from different channels are processed consistently through a unified interface and data format standardization technology to ensure the integrity and accuracy of data input.

[0042] Furthermore, the customer data after cleaning and formatting operations and the extracted feature data related to after-sales service are combined to obtain the score after data fusion, including the following specific contents: First, the data collected from social media, CRM system and customer feedback are cleaned, and the cleaning operation includes: deleting invalid data or duplicate data, filling missing values ​​and eliminating outliers. The invalid data is due to the noise generated during the data collection process, such as meaningless symbols, invalid entries generated by the system or incomplete records; duplicate data comes from the intersection between different data sources, such as duplicate information left by the same customer in social media and CRM system respectively. The data is marked and filtered by deduplication algorithm. For the processing of missing values ​​in the data, a variety of filling strategies are adopted. For numerical data, the average value or median is used for filling; for categorical data, the nearest neighbor algorithm is used for prediction and filling.

[0043] After cleaning, the data is formatted to convert data from multiple sources into a standardized format. Natural language processing technology is used to convert unstructured text data into analyzable features.

[0044] Based on text analysis technology, feature data related to after-sales service is extracted, and sentiment analysis algorithms are used to classify customer comments into positive, neutral or negative sentiments. The sentiment classification helps to identify the customer's satisfaction level with the current after-sales service. In addition, word frequency analysis and keyword extraction algorithms are used to identify words in the comments that are closely related to after-sales service, and semantic analysis technology is used to associate them with specific customer demand scenarios.

[0045] The extracted after-sales service feature data is combined with the data collected from social media, CRM system and customer feedback to establish a data fusion model, associate multi-source data and break the data silos. The formula of the data fusion model is:

[0046]

[0047] Where: D is the score after data fusion, S i is the i-th piece of social media data; T is the extracted after-sales service feature data; n is the number of social media data currently being processed; N S is the total amount of social media data; t is the current time point; t i is the timestamp of the ith piece of social media data; α is the weight coefficient of social media data in the overall fusion; γ is the smoothing factor of social media data, which controls the total amount of social media data N S The adjustment coefficient for the score after data fusion; β is the first time decay rate parameter; C j is the jth CRM system data; m is the number of data in the CRM system; C avg It is the average value of all relevant data feature values ​​in the CRM system, which is used to normalize and adjust the feature values ​​of CRM data; j is the timestamp of the CRM system data; δ is the weight coefficient of the data in the CRM system in the overall fusion; λ is the second time decay rate parameter; p is the number of customer feedback data items currently in use; F k is the k-th customer feedback data; N F is the total amount of customer feedback data; ζ is the smoothing factor of customer feedback data, which controls the total amount of customer feedback data N F The adjustment coefficient that affects the score after data fusion; η is the weight coefficient of customer feedback data in the overall fusion.

[0048] Furthermore, based on the output result of the data fusion model, an after-sales service demand prediction model is established to predict the customer's future after-sales service demand, including the following specific contents: using the output result of the data fusion model as input data, an after-sales service demand prediction model is established to predict the customer's future after-sales service demand, and the formula of the after-sales service demand prediction model is:

[0049]

[0050] Where: P(s) is the customer's after-sales service demand at the future time point s; D is the score after data fusion calculated by the data fusion model; s is the predicted time point; s0 is the current time; φ is the time decay rate; r is the total number of complaints submitted by customers; s k The time when the customer submits the complaint; ΔD k is the demand fluctuation, ΔD k satisfy in, is the time point s k The score after data fusion is is the time point s k The score after the data fusion at the previous time point.

[0051] Furthermore, according to the after-sales service demand, the allocation of after-sales service resources is optimized, including the following specific contents: based on the after-sales service demand prediction model, taking the after-sales service demand prediction as input, by optimizing the objective function, the goal of maximizing resource utilization and minimizing customer waiting time is achieved, and the formula of the objective function is:

[0052]

[0053] Where: M is the total number of after-sales service sites; W i is the average waiting time of customers at the i-th service station; P i is the after-sales service demand of the i-th service station; is the average waiting time of customers at the i-th service station W i The after-sales service demand of the i-th service station P i relationship; R i is the amount of resources allocated to the i-th service site; is the amount of resources R allocated to the i-th service site i The matching degree of after-sales service demand with the i-th service station; μ is the balance factor, which controls the balance between minimizing waiting time and resource matching; at the same time, in order to ensure that the total amount of resources allocated to each service station does not exceed the total amount of enterprise resources R total , to constrain the total amount of resources, the constraint formula is: And R iSatisfy R i ≥0, The objective function comprehensively considers the total amount of resources of the enterprise and the needs of each site. First, it takes the after-sales service demand provided by the demand forecasting model as the core, combines the resource allocation needs of each service site, and dynamically adjusts the allocation strategy. At the same time, it considers the customer waiting time and resource matching deviation, ensuring that the number of allocated resources can meet the site needs and avoid resource waste or shortage, and effectively reducing the waiting time of customers caused by unreasonable resource allocation. In addition, the total amount of resource constraints and non-negative constraints are added to ensure that the optimization plan is feasible under realistic constraints, and to achieve a balance between resource utilization and service efficiency through balancing factors.

[0054] Furthermore, customer demand trends and resource allocation plans are presented graphically, including the following specific contents: by integrating the output results of the after-sales service demand forecasting model and the optimization objective function, a variety of visualization methods are used to display information, including line charts, bar charts, pie charts, heat maps and interactive dashboards.

[0055] Step Y1: graphically display the customer demand trend, use a time series line chart to show the changing trend of customer demand in different time periods, and combine the forecast results to show the fluctuation range and growth point of future demand, so as to help enterprises predict the service peak period in advance;

[0056] Step Y2: Display resource allocation suggestions in a bar chart or stacked chart, specifying the resource allocation quantity for each service site, and marking the resource utilization rate with color coding to facilitate the discovery of potential problems in resource allocation;

[0057] Step Y3, providing a regional heat map function to display the spatial distribution of customer demand based on geographic location, helping enterprises identify demand hotspots and adjust resource layout;

[0058] Step Y5: In order to improve the user experience, a dynamic interactive function is also provided. Users can view detailed data of specific sites, such as customer waiting time, demand satisfaction rate and remaining resources, by clicking, dragging or filtering operations; and integrate key indicators and trend information into analysis results with graphics and texts to generate customized reports, so that decision makers can quickly obtain a global overview or conduct in-depth analysis of specific issues.

[0059] Through the above steps, enterprises can fully understand the current situation and future trends of customer needs and resource allocation, improve decision-making efficiency, optimize after-sales service processes, and ultimately enhance customer satisfaction and enterprise operational efficiency.

[0060] The present invention collects customer data from social media, CRM systems and customer feedback, and uses a variety of technical means to clean, format, extract features and fuse the data, establishes an after-sales service demand prediction model to predict future customer demand change trends, and then optimizes after-sales service resource allocation based on the prediction results. At the same time, it displays customer demand trends and resource allocation plans in a graphical manner, which can effectively integrate multi-source data, break data silos, deeply mine customer behavior and social data, predict after-sales service demand change trends, deploy resources in advance, reduce customer waiting time, improve customer satisfaction, and reduce customer churn risks; maximize resource utilization and minimize customer waiting time by optimizing the objective function, and consider a variety of realistic constraints to ensure the feasibility of the plan; graphical display facilitates enterprises to fully understand customer needs and resource allocation, improve decision-making efficiency, optimize after-sales service processes, and enhance enterprise operational efficiency.

[0061] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0062] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data analysis model based on SCRM, characterized in that: The steps include: Step S1, collecting customer data from social media, CRM system and customer feedback; Step S2, combining the customer data after cleaning and formatting operations and extracting the after-sales service related feature data to obtain the score D after data fusion, the contribution value of social media customer data to the fusion score is Among them, S i is the i-th piece of social media data, T is the extracted after-sales service feature data, n is the number of social media data currently being processed, N S is the total amount of social media data, t is the current time point, and t i is the timestamp of the ith piece of social media data, α is the weight coefficient of social media data in the overall fusion, γ is the smoothing factor of social media data, and β is the first time decay rate parameter; Step S3, based on the score after the data fusion, establish an after-sales service demand prediction model to predict the customer's future after-sales service demand. The formula of the after-sales service demand prediction model is: Where: P(s) is the customer's after-sales service demand at the future time point s; D is the score after data fusion calculated by the data fusion model; s is the predicted time point; s0 is the current time; φ is the time decay rate; r is the total number of complaints submitted by customers; s k The time when the customer submits the complaint; ΔD k is the demand fluctuation, ΔD k satisfy Among them, D sk is the time point s k The score after data fusion is is the time point s k The score after data fusion at the previous time point; Step S4, optimizing the allocation of after-sales service resources according to the demand for after-sales service; Step S5, displaying customer demand trends and resource allocation plans in a graphical manner.

2. According to the SCRM-based data analysis model of claim 1, it is characterized in that , the contribution of CRM system customer data to the fusion score is Among them, C j is the jth CRM system data, m is the number of data in the CRM system; C avg is the average value of all relevant data feature values ​​in the CRM system, t j is the timestamp of the CRM system data, δ is the weight coefficient of the data in the CRM system in the overall fusion, and λ is the second time decay rate parameter; the contribution of customer feedback data to the fusion score is Among them, p is the number of customer feedback data items currently used, F k is the kth customer feedback data, N F is the total amount of customer feedback data, ζ is the smoothing factor of customer feedback data, and controls the total amount of customer feedback data N F The adjustment coefficient that affects the score after data fusion, η is the weight coefficient of customer feedback data in the overall fusion.

3. The SCRM-based data analysis model according to claim 1 is characterized in that ,Based on the after-sales service demand prediction model, the after-sales service demand forecast is used as input, and the goal of maximizing resource utilization and minimizing customer waiting time is achieved by optimizing the objective function. The formula of the objective function is: Where: M is the total number of after-sales service sites; W i is the average waiting time of customers at the i-th service station; P i is the after-sales service demand of the i-th service station; is the average waiting time of customers at the i-th service station W i The after-sales service demand P of the i-th service station i relationship; R i is the amount of resources allocated to the i-th service site; is the amount of resources R allocated to the i-th service site i The matching degree of after-sales service demand with the i-th service station; μ is the balance factor, which controls the balance between minimizing waiting time and resource matching; at the same time, in order to ensure that the total amount of resources allocated to each service station does not exceed the total amount of enterprise resources R total , to constrain the total amount of resources, the constraint formula is: And R i Satisfy R i ≥0, 4. The SCRM-based data analysis model according to claim 1, characterized in that: In step S1, customer data from social media platforms are integrated when collecting data, including customer comment content, likes, sharing and reply interactive behaviors and feedback information related to product failures. Data connection with mainstream social media platforms is achieved through application programming interfaces and customer activities are dynamically monitored. Natural language processing is used to extract keywords and perform sentiment analysis on customer comments to identify the category and severity of fault problems.

5. The SCRM-based data analysis model according to claim 1 is characterized in that ,In the step S1, when collecting data, it connects with the CRM system to obtain the ,customer’s historical behavior data and structured information, including purchase records, ,transaction frequency, detailed information of purchased products, and interaction records with the ,customer service department.

6. The SCRM-based data analysis model according to claim 1, characterized in that ,In the step S1, the data sources of customer feedback channels are covered when collecting ,data, including the feedback information actively provided by customers through email, online customer ,service, telephone or questionnaire survey, and a distributed data collection ,architecture is adopted to process the data through a unified interface and data ,format standardization technology to ensure completeness and accuracy.

7. The SCRM-based data analysis model according to claim 1 is characterized in that In step S2, the cleaning operation includes deleting invalid data or duplicate data, filling missing values ​​and eliminating outliers. The invalid data is noise in the collection process, and the duplicate data comes from the intersection of different data sources. The data is deduplicated and different filling strategies are used for different types of data.

8. The SCRM-based data analysis model according to claim 1, characterized in that ,In the step S2, the formatting operation converts the collected ,customer data into a standardized format and uses natural language ,processing technology to convert the unstructured text data into analyzable ,features.

9. The SCRM-based data analysis model according to claim 1, characterized in that ,In the step S2, the extraction of after-sales service related ,feature data includes classifying the emotional tendency of customer comments through ,sentiment analysis algorithm, identifying related words using word frequency analysis and ,keyword extraction algorithm, and associating demand scenarios through semantic ,analysis technology.

Images