Customer relationship management system and method based on big data analysis

Abstract

The invention discloses a customer relationship management system and method based on big data analysis, and relates to the technical field of customer relationship management, and the method comprises the steps: collecting data from a plurality of sources, and carrying out the preprocessing; extracting features from the structured and unstructured data by using a data mining technology, and constructing a customer portrait template based on the features; based on the customer portrait template, monitoring new behavior data of the customer in real time, and adjusting the customer portrait template in real time; constructing a prediction model by using historical data, inputting features in the customer portrait template into the prediction model, and identifying customer demands; according to the prediction result and the personal preference in the historical behavior data, calculating the acceptance score of the customer for the solution; the solution is provided to the customer based on the customer's acceptance score of the solution.

Description

Technical Field

[0001] The present invention relates to the technical field of customer relationship management, and in particular to a customer relationship management system and method based on big data analysis. Background Art

[0002] In recent years, with the rapid development of information technology and the popularization of the Internet, Customer Relationship Management (CRM) systems have become an important tool for enterprises to improve customer satisfaction and enhance market competitiveness. Traditional CRM systems mainly rely on structured data, such as sales records and basic customer information, to identify customer behavior and needs through simple data analysis. However, with the advent of the big data era, the importance of unstructured data (social media interaction records, website visit records, and email interaction records) has become increasingly prominent. These unstructured data contain rich customer behavior and preference information, which can provide CRM systems with more comprehensive and in-depth insights. Therefore, how to effectively use these big data resources and improve the intelligence level of CRM systems has become one of the current research hotspots.

[0003] Although the existing customer relationship management system can help enterprises manage and analyze customer data to a certain extent, it still has obvious shortcomings in the following aspects. First, the data source is single. Most systems only rely on the internal database of the enterprise and ignore the value of external data sources. This makes the system unable to fully capture the multi-dimensional behaviors and preferences of customers, affecting the accuracy and completeness of customer portraits. Secondly, the data processing capacity is limited. When dealing with large-scale unstructured data, the existing system often faces the challenges of data cleaning, feature extraction and real-time processing, which makes it difficult for the system to respond to new behaviors of customers in a timely manner, reducing the timeliness and effectiveness of customer relationship management. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a customer relationship management system and method based on big data analysis to solve the problem of single data source. Most systems only rely on the internal database of the enterprise and ignore the value of external data sources. This makes the system unable to fully capture the multi-dimensional behaviors and preferences of customers, affecting the accuracy and completeness of customer portraits.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a customer relationship management method based on big data analysis, which includes:

[0008] Collect data from multiple sources and pre-process them;

[0009] Use data mining techniques to extract features from structured and unstructured data and build customer profile templates based on these features;

[0010] Based on the customer portrait template, monitor the customer's new behavior data in real time and adjust the customer portrait template in real time;

[0011] Use historical data to build a prediction model, input the features in the customer portrait template into the prediction model, and identify customer needs;

[0012] Calculate customer acceptance scores for solutions based on predictions and personal preferences from historical behavioral data;

[0013] Provide solutions to customers based on their acceptance scores of the solutions.

[0014] As a preferred solution of the customer relationship management method based on big data analysis described in the present invention, wherein: the data is collected from multiple sources and pre-processed, the specific steps are:

[0015] Collect customer data from internal corporate databases, social media platforms, website visit records, and email interaction records;

[0016] Use API interfaces to extract data from various data sources;

[0017] Use mean value filling method to fill missing values ​​in collected customer data;

[0018] For data with missing values ​​filled, the Z-score standardization method is used to standardize numerical data;

[0019] The standardized data is merged into the database table and each record is marked with a unique identifier, namely the customer ID.

[0020] As a preferred solution of the customer relationship management method based on big data analysis described in the present invention, wherein: the data mining technology is used to extract features from structured and unstructured data, and a customer portrait template is constructed based on these features. The specific steps are:

[0021] Extract data from database tables;

[0022] Feature extraction is performed on structured data in the database table. The expression is:

[0023]

[0024] Among them, TS represents the total consumption amount of the customer, It indicates accumulation from i=1 to i=n, ​​where n is the total number of transactions, i is the index variable, and TAi represents the amount of the i-th transaction; AS represents the average consumption amount, PF represents the purchase frequency, and TP represents the time period of the purchase frequency;

[0025] Feature extraction is performed on unstructured data in the database table. The expression is:

[0026] TF-IDF(t,d)=TF(t,d)×IDF(t);

[0027]

[0028] Among them, TF-IDF(t,d) represents the importance score of word t in document d, TF(t,d) represents the frequency of word t in document d, IDF(t) represents the inverse document frequency of word t in the entire document collection, log represents the logarithmic function, N represents the total number of documents, and df(t) represents the document frequency;

[0029] Use recursive feature elimination (RFE) to select the most important features;

[0030] Based on the extracted and selected features, use Pandas and JSON to build a customer portrait template;

[0031] Store the constructed customer portrait template in the database.

[0032] As a preferred solution of the customer relationship management method based on big data analysis described in the present invention, wherein: based on the customer portrait template, the new behavior data of the customer is monitored in real time, and the customer portrait template is adjusted in real time. The specific steps are:

[0033] Collect real-time data and pre-process the real-time data;

[0034] Based on the preprocessed data, extract the total consumption amount, average consumption amount, purchase frequency and customer profile related features from the structured data;

[0035] Extract TF-IDF value features for unstructured data;

[0036] Update the extracted features to the customer portrait template in real time, and dynamically adjust the customer portrait template. The expression is:

[0037]

[0038] Among them, DA represents the dynamic adjustment coefficient, Z represents the normalization constant, represents the sum of all features g from 1 to u, where u represents the number of features and w g (t) represents the weight of the g-th feature at time t, e represents the exponential function, and λ grepresents the decay rate of the g-th feature, s represents the integral variable, f represents the information filtering function, j represents the feature of the j-th information filtering function, m represents the number of features involved in information filtering, and v j (t) represents the actual observed value of the jth feature at time t, u j (t) represents the expected value of the j-th feature at time t, and ds represents the differential element in the integral.

[0039] As a preferred solution of the customer relationship management method based on big data analysis described in the present invention, wherein: the prediction model is constructed by using historical data, the features in the customer portrait template are input into the prediction model, and the customer's needs are identified. The specific steps are:

[0040] Use random forest as the basis of prediction model;

[0041] Input the features in the customer portrait template into the prediction model to identify customer needs. The prediction model expression is:

[0042]

[0043] Among them, PD represents the customer demand value, w g represents the weight of the g-th feature, x g It represents the value of the g-th feature in several intervals, and dx represents the differential element of the integral.

[0044] As a preferred solution of the customer relationship management method based on big data analysis described in the present invention, wherein: the customer's acceptance score for the solution is calculated based on the prediction results and personal preferences in the historical behavior data, and the specific steps are:

[0045] Collect customers’ historical behavior data from internal corporate databases, social media platforms, website visit records, and email interaction records, and calculate customers’ personal preferences based on the historical behavior data;

[0046]

[0047] Where PP represents the personal preference score, q represents the total number of different product types purchased by the customer, i and j represent index variables, and FP i represents the number of times a customer purchases the i-th product, T i represents the time interval between customers purchasing the i-th product, C represents the total number of different categories preferred by customers, PC j Indicates the customer's preference for the jth category, PT j represents the number of times a customer purchases in the jth category, and PE represents the total number of different interaction channels preferred by the customer;

[0048] The feedback scoring formula is constructed based on the prediction results and personal preferences. The expression is:

[0049]

[0050] Among them, PS represents the customer's acceptance score of the solution, log represents the logarithmic function, e represents the base of the natural logarithm, μ represents the parameter controlling the steepness of the Sigmoid function, PF represents the purchase frequency, and PF avg represents the average purchase frequency, cos represents the cosine function, and π represents the pi.

[0051] As a preferred solution of the customer relationship management method based on big data analysis according to the present invention, wherein: providing a solution to the customer based on the customer's acceptance score of the solution includes,

[0052] The value range of the customer's acceptance score PS of the solution is [0, 1];

[0053] Set the threshold YY;

[0054] When YY ≤ PS ≤ 1, it indicates that the customer has a high acceptance of the solution;

[0055] When 0 ≤ PS < YY, it indicates that the customer has a low acceptance of the solution;

[0056] When YY ≤ PS ≤ 1, the customer manager will regularly contact the customer to understand their needs, provide personalized suggestions and services, and the customer's problems and requests will be given priority;

[0057] When 0 ≤ PS < YY, provide targeted high discounts for low-acceptance customers, provide free delivery services for low-acceptance customers when the order amount meets the threshold, and provide point rewards for low-acceptance customers to encourage them to make another purchase.

[0058] In a second aspect, the present invention provides a customer relationship management system based on big data analysis, including,

[0059] A collection module, responsible for collecting customer data from multiple data sources and performing preprocessing;

[0060] A feature extraction module, responsible for extracting features from structured and unstructured data and constructing a customer portrait template based on these features;

[0061] A monitoring and adjustment module, responsible for real-time monitoring of the customer's new behavior data and dynamically adjusting the customer portrait template;

[0062] A prediction module, responsible for using historical data to construct a prediction model, identifying the customer's needs, and providing solutions;

[0063] The feedback module provides solutions to customers based on their acceptance scores and collects customer feedback to further optimize the system.

[0064] In a third aspect, an embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the customer relationship management method based on big data analysis as described in the first aspect of the present invention is implemented.

[0065] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the customer relationship management method based on big data analysis as described in the first aspect of the present invention.

[0066] The beneficial effects of the present invention are as follows: the present invention collects data from multiple sources, performs preprocessing, and uses data mining technology to extract features from structured and unstructured data to build a customer portrait template. At the same time, the system can monitor the customer's new behavior data in real time and dynamically adjust the customer portrait template to ensure the real-time and accuracy of the customer portrait. By using historical data to build a prediction model, the system can identify customer needs and calculate the customer's acceptance score for the solution based on the prediction results and personal preferences in the historical behavior data. Finally, the system provides personalized solutions based on the customer's acceptance score for the solution to improve customer satisfaction and loyalty. The present invention solves the problems of existing customer relationship management systems such as single data source, limited data processing capabilities, low prediction model accuracy, and insufficient personalized recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0068] Figure 1 This is a flow chart of the customer relationship management method based on big data analysis in Example 1.

[0069] Figure 2 This is a flow chart of the customer portrait template in Example 1. DETAILED DESCRIPTION

[0070] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0071] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0072] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0073] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a customer relationship management method based on big data analysis, comprising the following steps:

[0074] S1 collects data from multiple sources and performs preprocessing;

[0075] Collect customer data from internal corporate databases, social media platforms, website visit records, and email interaction records. Customer data includes sales records, social media interaction records, and website visit records;

[0076] Use API interfaces to extract data from various data sources. For data sources that do not have APIs, use web crawler technology to crawl data;

[0077] Use mean value filling method to fill missing values ​​in collected customer data;

[0078] For data with missing values ​​filled, the Z-score standardization method is used to standardize numerical data;

[0079] Merge the standardized data into the database table and give each record a unique identifier, i.e., the customer ID;

[0080] Internal enterprise databases refer to ERP systems and CRM systems;

[0081] Social media platforms refer to Weibo and WeChat public accounts

[0082] S2 uses data mining technology to extract features from structured and unstructured data and build customer profile templates based on these features;

[0083] Extract data from database tables;

[0084] Feature extraction is performed on structured data in the database table. The expression is:

[0085]

[0086] Among them, TS represents the total consumption amount of the customer, It indicates accumulation from i=1 to i=n, ​​where n is the total number of transactions, i is the index variable, and TA i represents the amount of the i-th transaction; AS represents the average consumption amount, PF represents the purchase frequency, and TP represents the time period of the purchase frequency;

[0087] By extracting features from unstructured data in the database table and combining structured and unstructured data, we can fully capture the multi-dimensional behaviors and preferences of customers and improve the accuracy and completeness of customer portraits. The expression is:

[0088] TF-IDF(t,d)=TF(t,d)×IDF(t);

[0089]

[0090] Among them, TF-IDF(t,d) represents the importance score of word t in document d, TF(t,d) represents the frequency of word t in document d, IDF(t) represents the inverse document frequency of word t in the entire document collection, log represents the logarithmic function, N represents the total number of documents, and df(t) represents the document frequency;

[0091] Use recursive feature elimination (RFE) to select the most important features. The recursive feature elimination (RFE) method can automatically select the most important features, reduce feature dimensions, and improve model training efficiency and prediction accuracy.

[0092] Based on the extracted and selected features, Pandas and JSON are used to build customer portrait templates. Pandas is used for data processing and formatting, and JSON is used to store and transmit data.

[0093] Storing the constructed customer portrait template in the database ensures the security and persistence of the data and facilitates subsequent query and analysis.

[0094] S3 monitors customers’ new behavior data in real time based on customer portrait templates, and adjusts customer portrait templates in real time;

[0095] Collect real-time data and pre-process the real-time data;

[0096] Based on the preprocessed data, extract the total consumption amount, average consumption amount, purchase frequency and customer profile related features from the structured data;

[0097] Extract TF-IDF value features for unstructured data;

[0098] Update the extracted features to the customer portrait template in real time, and dynamically adjust the customer portrait template. The expression is:

[0099]

[0100] Among them, DA represents the dynamic adjustment coefficient, Z represents the normalization constant, represents the sum of all features g from 1 to u, where u represents the number of features and w g (t) represents the weight of the g-th feature at time t, e represents the exponential function, and λ g represents the decay rate of the g-th feature, s represents the integral variable, f represents the information filtering function, j represents the feature of the j-th information filtering function, m represents the number of features involved in information filtering, and v j (t) represents the actual observed value of the jth feature at time t, u j (t) represents the expected value of the jth feature at time t, and ds represents the differential element in the integral;

[0101] Information filtering function f, the calculation formula is:

[0102]

[0103] Where y represents the label, Y represents the target variable, p(x,y) represents the probability that the feature value x and the label y appear at the same time, p(x) represents the probability that the feature value x appears, p(y) represents the probability that the label y appears, and log represents the natural logarithm;

[0104] The value range of the dynamic adjustment coefficient DA is from 0 to positive infinity. When DA>1, it means that the customer behavior has increased; when DA<1, it means that the customer behavior has decreased; when DA=1, it means that the customer behavior remains unchanged.

[0105] S4 uses historical data to build a prediction model, inputs the features in the customer portrait template into the prediction model, and identifies customer needs;

[0106] Use random forest as the basis of prediction model;

[0107] Inputting the features in the customer portrait template into the prediction model, and inputting multiple features such as total consumption amount, average consumption amount, purchase frequency and TF-IDF value into the prediction model can fully capture the multi-dimensional behavior and preferences of customers, improve the accuracy of demand forecasting, and identify customer needs. The prediction model expression is:

[0108]

[0109] Among them, PD represents the customer demand value, w g represents the weight of the g-th feature, xg represents the value of the g-th feature in several intervals, and dx represents the differential element of the integral;

[0110] The value range of the customer demand value PD is from 0 to 1. When PD = 1, it means that the customer's demand is very high. When PD = 0, it means that the customer's demand is very low. When PD = 0.5, it means that the customer's demand is at a medium level.

[0111] S5 calculates the customer's acceptance score for the solution based on the prediction results and personal preferences in historical behavior data;

[0112] Collect customers’ historical behavior data from internal corporate databases, social media platforms, website visit records, and email interaction records, and calculate customers’ personal preferences based on the historical behavior data;

[0113]

[0114] Where PP represents the personal preference score, q represents the total number of different product types purchased by the customer, i and j represent index variables, and FP i represents the number of times a customer purchases the i-th product, T i represents the time interval between customers purchasing the i-th product, C represents the total number of different categories preferred by customers, PC j Indicates the customer's preference for the jth category, PT j represents the number of times a customer purchases in the jth category, and PE represents the total number of different interaction channels preferred by the customer;

[0115] A feedback scoring formula is constructed based on the prediction results and personal preferences to evaluate the customer's acceptance of the solution. The expression is:

[0116]

[0117] Among them, PS represents the customer's acceptance score of the solution, log represents the logarithmic function, e represents the base of the natural logarithm, μ represents the parameter that controls the steepness of the Sigmoid function, PF represents the purchase frequency, and PF avg represents the average purchase frequency, cos represents the cosine function, and π represents the pi ratio. The logarithmic function, sigmoid function, and cosine function are used to perform weighted processing on data of different dimensions to ensure the rationality and accuracy of the ratings.

[0118] The formula takes into account prediction results, personal preferences, purchase frequency and interaction channels to more accurately assess customer acceptance;

[0119] The total number of PEs for different preferred interaction channels, including social media, email, website visits, and telephone consultations.

[0120] S6 provides a solution to the customer based on the customer's acceptance score of the solution;

[0121] The value range of the acceptance score PS of the customer for the solution is [0, 1];

[0122] Set the threshold YY;

[0123] When YY ≤ PS ≤ 1, it means that the customer has a high acceptance of the solution;

[0124] When 0 ≤ PS < YY, it means that the customer has a low acceptance of the solution;

[0125] When YY ≤ PS ≤ 1, the account manager will regularly contact the customer to understand their needs, provide personalized advice and services, the customer's problems and requests will be given priority, ensure a quick response, and generate coupons based on the customer's purchase history;

[0126] When 0 ≤ PS < YY, provide targeted high discounts for low-acceptance customers, including a 20% discount on specific products. When the order amount meets the threshold, provide free delivery services for low-acceptance customers. When the order amount exceeds 100 yuan, enjoy free delivery services. Provide point rewards for low-acceptance customers to encourage them to make another purchase, including providing additional point rewards every time they shop. Customers can use the points to exchange for gifts or deduct part of the order amount.

[0127] This embodiment also provides a customer relationship management system based on big data analysis, including:

[0128] A collection module, responsible for collecting customer data from multiple data sources and performing preprocessing;

[0129] A feature extraction module, responsible for extracting features from structured and unstructured data and constructing a customer portrait template based on these features;

[0130] A monitoring and adjustment module, responsible for real-time monitoring of the customer's new behavior data and dynamically adjusting the customer portrait template;

[0131] A prediction module, responsible for using historical data to construct a prediction model, identifying the customer's needs, and providing solutions;

[0132] A feedback module, providing solutions to the customer according to the customer's acceptance score and collecting the customer's feedback to further optimize the system.

[0133] This embodiment also provides a computer device, which is suitable for the case of a customer relationship management method based on big data analysis, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the customer relationship management method based on big data analysis as proposed in the above embodiment.

[0134] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0135] The present embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the customer relationship management method based on big data analysis proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0136] In summary, the present invention collects data from multiple sources and performs preprocessing, uses data mining technology to extract features from structured and unstructured data, and constructs a customer portrait template. At the same time, the system can monitor the customer's new behavior data in real time and dynamically adjust the customer portrait template to ensure the real-time and accuracy of the customer portrait. By using historical data to build a prediction model, the system can identify customer needs and calculate the customer's acceptance score for the solution based on the prediction results and personal preferences in the historical behavior data. Finally, the system provides personalized solutions based on the customer's acceptance score for the solution, thereby improving customer satisfaction and loyalty. The present invention solves the problems of existing customer relationship management systems, such as a single data source, limited data processing capabilities, low prediction model accuracy, and insufficient personalized recommendations.

[0137] Example 2, referring to Table 1, is the second example of the present invention. To further verify the technical solution of the present invention, experimental simulation data of a customer relationship management method based on big data analysis is provided.

[0138] In order to verify the effectiveness and superiority of the customer relationship management method based on big data analysis, this study selected the customer data of a large retail enterprise as a sample, aiming to improve customer satisfaction and loyalty through this method. During the experiment, customer data was first collected from multiple sources, including the enterprise's internal ERP system and CRM system, social media platforms such as Weibo and WeChat public accounts, as well as official websites and email interaction records. In view of the characteristics of different data sources, a combination of API interface calls and web crawler technology was adopted to ensure the comprehensiveness and accuracy of data collection.

[0139] In the data preprocessing stage, the collected raw data were cleaned, including removing duplicate records, correcting erroneous information, and filling missing values. Specifically, the mean value filling method was used to handle missing values ​​to ensure the integrity and consistency of the data. Afterwards, the Z-score standardization method was used to standardize the numerical data to ensure the comparability between data of different dimensions. The preprocessed data were merged into a unified database table, and each record was given a unique identifier - customer ID, to facilitate subsequent analysis and management.

[0140] Next, we extracted key features from structured and unstructured data using data mining techniques and built a customer profile template. For structured data, we calculated the customer’s total spending amount, average spending amount, and purchase frequency, and applied the TF-IDF algorithm to process unstructured text data to quantify customer interests and preferences. We used the recursive feature elimination (RFE) method to further screen out the most influential features, improving the model’s training efficiency and prediction accuracy.

[0141] In order to maintain the timeliness and accuracy of customer portraits, the system also implements a real-time monitoring and dynamic adjustment mechanism for new customer behavior data. Whenever a new transaction occurs or social media interaction occurs, the system will immediately update the relevant features and reflect the changing trend of customer behavior by calculating the dynamic adjustment coefficient. In addition, the prediction model built based on historical data can effectively identify potential customer needs, so that measures can be taken in advance to meet customer expectations.

[0142] Finally, by combining the prediction results with personal preferences in historical behavioral data, we calculated customers’ acceptance scores for the solution and developed differentiated marketing strategies accordingly. For customers with high acceptance, we provided personalized advice and services. For customers with low acceptance, we provided more preferential policies and support in order to improve their perception and attitude toward the brand.

[0143] The details are shown in Table 1 below:

[0144] Table 1 Experimental record table

[0145]

[0146] Through in-depth analysis of the above table data, it can be clearly seen that the customer relationship management method provided by the present invention has shown significant advantages in improving customer satisfaction and loyalty; first, in terms of customer portrait construction, through the comprehensive use of structured data (such as total consumption amount, average consumption amount and purchase frequency) and unstructured data (such as TF-IDF value), it can not only fully capture the customer's consumption habits and interest preferences, but also effectively improve the accuracy and completeness of the customer portrait; for example, customer E has the highest total consumption amount, and its TF-IDF value also reaches the maximum, which shows that the customer's activity on social media and brand loyalty are very high;

[0147] Secondly, the application of the dynamic adjustment mechanism enables customer portraits to reflect changes in customer behavior in a timely manner, which is crucial for formulating accurate marketing strategies. For example, the dynamic adjustment coefficient (DA) of customer A is close to 1, which means that his consumption behavior is relatively stable; while the DA value of customer D is less than 1, which indicates that we need to pay attention to changes in the customer's behavior and adjust the service strategy in a timely manner to prevent customer churn.

[0148] Furthermore, the introduction of the prediction model has greatly improved our ability to predict customer needs. Taking customer C as an example, both his personal preference score (PP) and customer acceptance score (PS) of the solution are at a high level, which shows that our prediction model can accurately identify such high-value customers and provide them with more considerate services, thereby promoting the improvement of customer conversion rate.

[0149] In summary, the method proposed in the present invention not only overcomes the problems of data isolation and one-sided analysis in traditional customer relationship management, but also realizes the deep mining and intelligent application of customer data through technological innovation, bringing tangible commercial value to enterprises; compared with the existing customer relationship management system, this method has excellent performance in data processing capability, depth of customer insight and flexibility of marketing strategy, and has strong practicality and promotion prospects.

[0150] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A customer relationship management method based on big data analysis, characterized in that: include, Collect data from multiple sources and pre-process them; Use data mining techniques to extract features from structured and unstructured data and build customer profile templates based on these features; Based on the customer portrait template, monitor the customer's new behavior data in real time and adjust the customer portrait template in real time; Use historical data to build a prediction model, input the features in the customer portrait template into the prediction model, and identify customer needs; Calculate customer acceptance scores for solutions based on predictions and personal preferences from historical behavioral data; Provide solutions to customers based on their acceptance scores of the solutions.

2. The customer relationship management method based on big data analysis according to claim 1, characterized in that: The data is collected from multiple sources and preprocessed. The specific steps are: Collect customer data from internal corporate databases, social media platforms, website visit records, and email interaction records; Use API interfaces to extract data from various data sources; Use mean value filling method to fill missing values ​​in collected customer data; For data with missing values ​​filled, the Z-score standardization method is used to standardize numerical data; The standardized data is merged into the database table and each record is marked with a unique identifier, namely the customer ID.

3. The customer relationship management method based on big data analysis according to claim 2, characterized in that: The data mining technology is used to extract features from structured and unstructured data, and a customer portrait template is constructed based on these features. The specific steps are: Extract data from database tables; Feature extraction is performed on structured data in the database table. The expression is: Among them, TS represents the total consumption amount of the customer, It indicates accumulation from i=1 to i=n, ​​where n is the total number of transactions, i is the index variable, and TA i represents the amount of the i-th transaction; AS represents the average consumption amount, PF represents the purchase frequency, and TP represents the time period of the purchase frequency; Feature extraction is performed on unstructured data in the database table. The expression is: TF-IDF(t,d)=TF(t,d)×IDF(t); Among them, TF-IDF(t,d) represents the importance score of word t in document d, TF(t,d) represents the frequency of word t in document d, IDF(t) represents the inverse document frequency of word t in the entire document collection, log represents the logarithmic function, N represents the total number of documents, and df(t) represents the document frequency; Use recursive feature elimination (RFE) to select the most important features; Based on the extracted and selected features, use Pandas and JSON to build a customer portrait template; Store the constructed customer portrait template in the database.

4. The customer relationship management method based on big data analysis according to claim 3, characterized in that: Based on the customer portrait template, the new behavior data of the customer is monitored in real time, and the customer portrait template is adjusted in real time. The specific steps are: Collect real-time data and pre-process the real-time data; Based on the preprocessed data, extract the total consumption amount, average consumption amount, purchase frequency and customer profile related features from the structured data; Extract TF-IDF value features for unstructured data; Update the extracted features to the customer portrait template in real time, and dynamically adjust the customer portrait template. The expression is: Among them, DA represents the dynamic adjustment coefficient, Z represents the normalization constant, represents the sum of all features g from 1 to u, where u represents the number of features and w g (t) represents the weight of the g-th feature at time t, e represents the exponential function, and λ g represents the decay rate of the g-th feature, s represents the integral variable, f represents the information filtering function, j represents the feature of the j-th information filtering function, m represents the number of features involved in information filtering, and v j (t) represents the actual observed value of the jth feature at time t, u j (t) represents the expected value of the j-th feature at time t, and ds represents the differential element in the integral.

5. The customer relationship management method based on big data analysis according to claim 4, characterized in that: The prediction model is constructed by using historical data, and the features in the customer portrait template are input into the prediction model to identify customer needs. The specific steps are as follows: Use random forest as the basis of prediction model; Input the features in the customer portrait template into the prediction model to identify customer needs. The prediction model expression is: Among them, PD represents the customer demand value, w g represents the weight of the g-th feature, x g It represents the value of the g-th feature in several intervals, and dx represents the differential element of the integral.

6. The customer relationship management method based on big data analysis according to claim 5, characterized in that: Calculating the acceptance score of the customer for the solution based on the prediction result and the personal preferences in the historical behavior data. The specific steps are as follows: Collect the historical behavior data of the customer from the enterprise internal database, social media platforms, website access records, and email interaction records, and calculate the personal preference degree of the customer based on the historical behavior data; Where PP represents the personal preference score, q represents the total number of different product types purchased by the customer, i and j represent index variables, and FP i represents the number of times a customer purchases the i-th product, T i represents the time interval between customers purchasing the i-th product, C represents the total number of different categories preferred by customers, PC j Indicates the customer's preference for the jth category, PT j represents the number of times a customer purchases in the jth category, and PE represents the total number of different interaction channels preferred by the customer; Construct a feedback score formula based on the prediction result and personal preferences. The expression is: Among them, PS represents the customer's acceptance score of the solution, log represents the logarithmic function, e represents the base of the natural logarithm, μ represents the parameter that controls the steepness of the Sigmoid function, FP represents the purchase frequency, and PF represents the purchase frequency. avg represents the average purchase frequency, cos represents the cosine function, and π represents the pi.

7. The customer relationship management method based on big data analysis according to claim 6, characterized in that: Providing the solution to the customer based on the acceptance score of the customer for the solution, including: The value range of the acceptance score PS of the customer for the solution is [0, 1]; Set the threshold YY; When YY ≤ PS ≤ 1, it indicates that the customer has a high acceptance of the solution; When 0 ≤ PS < YY, it indicates that the customer has a low acceptance of the solution; When YY ≤ PS ≤ 1, the customer manager will regularly contact the customer to understand their needs, provide personalized suggestions and services, and the problems and requests of the customer will be given priority; When 0 ≤ PS < YY, provide targeted high discounts for customers with low acceptance. When the order amount meets the threshold, provide free delivery services for customers with low acceptance, and provide points rewards for customers with low acceptance to encourage them to make another purchase.

8. A customer relationship management system based on big data analysis, based on the customer relationship management method based on big data analysis according to any one of claims 1 to 7, characterized in that: Including: A collection module, responsible for collecting customer data from multiple data sources and performing preprocessing; A feature extraction module, responsible for extracting features from structured and unstructured data and constructing a customer portrait template based on these features; A monitoring and adjustment module, responsible for real-time monitoring of the new behavior data of the customer and dynamically adjusting the customer portrait template; A prediction module, responsible for constructing a prediction model using historical data, identifying the needs of the customer, and providing solutions; A feedback module, providing solutions to the customer according to the acceptance score of the customer, and collecting the feedback of the customer to further optimize the system.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the customer relationship management method based on big data analysis according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the customer relationship management method based on big data analysis according to any one of claims 1 to 7.

Images