Personalized service recommendation method and device, equipment and storage medium

Abstract

The invention discloses a personalized service recommendation method, device and equipment and a storage medium, and relates to the technical field of big data, and the method comprises the steps: collecting feedback data of clients in different channels in a unified manner through a platform, and determining a feedback sentiment analysis result, a customer satisfaction score value, a potential problem, a customer life cycle stage, point rewards, exclusive preferences and comprehensive behavior characteristics according to customer feedback data, and comprehensively generating a personalized service recommendation scheme. The feedback data on different channels are collected through the unified platform, and the validity of the feedback data is ensured; real-time analysis is carried out according to collected feedback data, analysis is carried out in the aspects of feedback emotion, customer satisfaction, potential problem recognition, life cycle stages, point rewards, exclusive preferences and the like, the real emotion and requirements of customers are accurately recognized, and the accuracy and reliability of analysis results are ensured; and personalized service recommendation is provided through comprehensive analysis of behavior characteristics, so that the customer satisfaction is improved.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular to a personalized service recommendation method, apparatus, device, and storage medium. Background Technology

[0002] Customer feedback and satisfaction management has become a crucial aspect of modern enterprises. To remain competitive and meet customer needs, many companies employ various methods to collect and analyze customer feedback in order to better understand customer needs and improve service quality.

[0003] Existing customer feedback and satisfaction management methods have shortcomings. Companies typically collect customer feedback through independent channels (such as telephone, email, and social media). Feedback data from each channel is relatively independent, lacking unified and comprehensive analysis, making it difficult to fully reflect the overall customer experience and resulting in insufficient representativeness and validity of the feedback data. Current customer feedback analysis relies heavily on static data, lacking real-time and dynamic adjustment capabilities, making it difficult to accurately identify customers' true emotions and needs, leading to low customer satisfaction with service recommendations. Summary of the Invention

[0004] The main objective of this application is to provide a personalized service recommendation method, apparatus, device, and storage medium, aiming to solve the technical problem that existing methods suffer from insufficient validity of feedback data and lack of real-time dynamic adjustment capabilities, resulting in low customer satisfaction with service recommendations.

[0005] To achieve the above objectives, this application proposes a personalized service recommendation method, the method comprising:

[0006] Customer feedback data from various channels is collected uniformly through a pre-set platform;

[0007] Based on the customer feedback data, feedback sentiment analysis is performed to obtain feedback sentiment analysis results;

[0008] Based on the customer feedback data, a customer satisfaction score is obtained.

[0009] Based on the feedback sentiment analysis results and the customer satisfaction score, potential problems are identified to obtain potential problem identification results.

[0010] The customer lifecycle stage is determined based on the feedback sentiment analysis results;

[0011] The points-based reward is determined based on the customer lifecycle stage and the customer satisfaction score.

[0012] Exclusive offers are determined based on the customer lifecycle stage and the points rewards;

[0013] A comprehensive analysis is conducted based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward to obtain comprehensive behavioral characteristics;

[0014] Based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, a personalized service recommendation plan is determined and pushed to the user.

[0015] In one embodiment, the step of uniformly collecting customer feedback data from multiple different channels through a preset platform includes:

[0016] Customer feedback data from various channels is collected uniformly through a pre-set platform;

[0017] The channel weights for the various channels are determined based on customer coverage, feedback quality, customer preferences, and response speed.

[0018] The feedback data from the various channels are weighted and summed according to the channel weights to obtain the channel weight summation result.

[0019] The feedback influencing factors are determined based on feedback timeliness, customer lifecycle stage, changes in the market environment, and historical feedback data.

[0020] The feedback impact value is determined based on the feedback impact factor and the life cycle stage corresponding to the previous time of the preset time;

[0021] The summation of the channel weights and the feedback impact value are summed to obtain customer feedback data.

[0022] In one embodiment, the step of performing feedback sentiment analysis based on the customer feedback data to obtain feedback sentiment analysis results includes:

[0023] The feedback weight of each customer's feedback data is determined based on the feedback importance weight, time relevance weight, and feedback source weight.

[0024] Based on the feedback weights, the sentiment scores of each customer's feedback data in the customer feedback data are weighted and summed to obtain the feedback weight summation result;

[0025] The customer points reward value for each customer is determined based on the customer sentiment influence factor and the customer points reward.

[0026] The feedback sentiment analysis result for each customer in the customer feedback data is determined based on the summation result of the feedback weights and the customer points reward value.

[0027] In one embodiment, the step of scoring customer satisfaction based on the customer feedback data to obtain a customer satisfaction score includes:

[0028] The satisfaction scores of each customer in the customer feedback data are weighted and summed for each rating based on the satisfaction rating weights to obtain the satisfaction summation result.

[0029] The satisfaction impact value of each customer is determined based on the satisfaction impact factors and the feedback sentiment analysis results of each customer.

[0030] The customer satisfaction score for each customer in the customer feedback data is determined based on the summation of the satisfaction scores and the impact value of the satisfaction scores.

[0031] In one embodiment, the step of identifying potential problems based on the feedback sentiment analysis results and the customer satisfaction score to obtain potential problem identification results includes:

[0032] The dynamic adjustment factor is obtained by adjusting the historical feedback sentiment score and historical customer satisfaction score data according to the adjustment coefficient.

[0033] Alternatively, adjustment coefficients can be set based on business objectives, and dynamic adjustment factors can be determined based on these adjustment coefficients.

[0034] Alternatively, machine learning can be used to predict historical feedback sentiment scores and historical customer satisfaction scores, and dynamic adjustment factors can be determined based on the prediction results.

[0035] Determine the difference between the feedback sentiment analysis result and the customer satisfaction score;

[0036] The potential problem identification result is determined based on the difference, the preset difference threshold, and the dynamic adjustment factor.

[0037] In one embodiment, the step of determining the customer lifecycle stage based on the feedback sentiment analysis results includes:

[0038] The linear correlation between historical feedback sentiment scores and historical life cycle stage data was determined based on the Pearson correlation coefficient, and the linear correlation was adjusted based on the adjustment coefficient to obtain the life cycle adjustment factor.

[0039] The customer activity level of each customer is determined based on online behavior data analysis, transaction data analysis, and interaction data analysis.

[0040] The customer lifecycle stage is determined based on the customer activity level, the feedback sentiment analysis results, the first activity threshold, the second activity threshold, the first feedback sentiment score threshold, the second feedback sentiment score threshold, and the lifecycle adjustment factor. The customer lifecycle stage includes the introduction stage, the growth stage, the maturity stage, and the decline stage. The first feedback sentiment score threshold is greater than or equal to the second feedback sentiment score threshold.

[0041] In one embodiment, after the step of determining a personalized service recommendation scheme based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, the method further includes:

[0042] The personalized service recommendation scheme is evaluated to obtain the evaluation results;

[0043] The effectiveness of the personalized service recommendation scheme will be determined based on the evaluation results.

[0044] If the personalized service recommendation scheme is invalid, the customer's potential problems will be adjusted according to the scheme adjustment coefficient and the evaluation results, and the adjusted personalized service recommendation scheme will be determined according to the adjusted customer's potential problems.

[0045] The points reward is adjusted according to the weighting coefficient to obtain the adjusted points reward;

[0046] The adjusted personalized service recommendation scheme is optimized based on the adjusted points reward system to obtain an optimized personalized service recommendation scheme, which is then pushed to the user.

[0047] Furthermore, to achieve the above objectives, this application also proposes a personalized service recommendation device, which includes:

[0048] The customer feedback data collection module is used to collect customer feedback data from various channels through a pre-set platform.

[0049] The feedback sentiment analysis module is used to perform feedback sentiment analysis based on the customer feedback data and obtain feedback sentiment analysis results.

[0050] The customer satisfaction rating module is used to score customer satisfaction based on the customer feedback data and obtain a customer satisfaction rating value.

[0051] The potential problem identification module is used to identify potential problems based on the feedback sentiment analysis results and the customer satisfaction score, and obtain potential problem identification results.

[0052] The customer lifecycle determination module is used to determine the customer lifecycle stage based on the feedback sentiment analysis results.

[0053] The points reward determination module is used to determine points rewards based on the customer lifecycle stage and the customer satisfaction score.

[0054] The exclusive offer determination module is used to determine exclusive offers based on the customer lifecycle stage and the points rewards.

[0055] The behavioral characteristic comprehensive analysis module is used to perform comprehensive analysis based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward to obtain comprehensive behavioral characteristics;

[0056] The personalized service recommendation module is used to determine a personalized service recommendation plan based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, and to push the personalized service recommendation plan to the user.

[0057] In addition, to achieve the above objectives, this application also proposes a personalized service recommendation device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the personalized service recommendation method as described above.

[0058] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and which, when executed by a processor, implements the steps of the personalized service recommendation method described above.

[0059] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the personalized service recommendation method described above.

[0060] This application provides a personalized service recommendation method. It collects customer feedback data from various channels through a pre-set platform. Based on this data, it determines feedback sentiment analysis results, customer satisfaction scores, potential problems, customer lifecycle stages, points rewards, exclusive offers, and comprehensive behavioral characteristics, and then generates a personalized service recommendation scheme. This application ensures the representativeness and validity of the feedback data by collecting it through a unified platform. Real-time analysis of the collected feedback data, focusing on feedback sentiment, customer satisfaction, potential problem identification, lifecycle stages, points rewards, and exclusive offers, accurately identifies customers' true emotions and needs, ensuring the accuracy and reliability of the analysis results. Through comprehensive behavioral characteristic analysis, it provides personalized service recommendations, thereby improving customer satisfaction. Attached Figure Description

[0061] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0062] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 This is a flowchart illustrating the personalized service recommendation method of this application in Implementation Example 1.

[0064] Figure 2 This is an interactive diagram illustrating the personalized service recommendation method used in this application.

[0065] Figure 3 This is a flowchart illustrating Embodiment 2 of the personalized service recommendation method for this application.

[0066] Figure 4 This is a flowchart illustrating Embodiment 3 of the personalized service recommendation method for this application.

[0067] Figure 5 An interactive diagram illustrating the optimization process of the personalized service recommendation method for this application;

[0068] Figure 6 This is a schematic diagram of the module structure of the personalized service recommendation device according to an embodiment of this application;

[0069] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the personalized service recommendation method in this application embodiment.

[0070] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0071] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0072] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0073] The main solution of this application embodiment is as follows: Customer feedback data from various channels is collected uniformly through a preset platform; feedback sentiment analysis is performed based on the customer feedback data to obtain feedback sentiment analysis results; customer satisfaction is scored based on the customer feedback data to obtain customer satisfaction score values; potential problems are identified based on the feedback sentiment analysis results and the customer satisfaction score values ​​to obtain potential problem identification results; customer lifecycle stages are determined based on the feedback sentiment analysis results; points rewards are determined based on the customer lifecycle stages and the customer satisfaction score values; exclusive offers are determined based on the customer lifecycle stages and the points rewards; comprehensive analysis is performed based on the feedback sentiment analysis results, the customer satisfaction score values, the potential problems, the customer lifecycle stages, and the points rewards to obtain comprehensive behavioral characteristics; a personalized service recommendation scheme is determined based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score values, the potential problems, the customer lifecycle stages, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, and the personalized service recommendation scheme is pushed to the user.

[0074] Due to shortcomings in existing customer feedback and satisfaction management methods, businesses typically collect customer feedback through independent channels (such as telephone, email, and social media). Feedback data from each channel is relatively independent, lacking unified and comprehensive analysis, making it difficult to fully reflect the overall customer experience and resulting in insufficient representativeness and validity of the feedback data. Existing customer feedback analysis relies heavily on static data, lacking real-time and dynamic adjustment capabilities, making it difficult to accurately identify customers' true emotions and needs, leading to low customer satisfaction with service recommendations.

[0075] This application provides a solution that collects customer feedback data from various channels through a pre-defined platform. Based on this data, it determines feedback sentiment analysis results, customer satisfaction scores, potential problems, customer lifecycle stages, points rewards, exclusive offers, and comprehensive behavioral characteristics, and then generates a personalized service recommendation plan. This application ensures the representativeness and validity of the feedback data by collecting it through a unified platform. Real-time analysis of the collected feedback data, focusing on aspects such as feedback sentiment, customer satisfaction, potential problem identification, lifecycle stages, points rewards, and exclusive offers, accurately identifies customers' true emotions and needs, ensuring the accuracy and reliability of the analysis results. Through comprehensive behavioral characteristic analysis, it provides personalized service recommendations, thereby improving customer satisfaction.

[0076] It should be noted that the executing entity of the method in this embodiment can be a computing service device with personalized service recommendation, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone; or it can be a personalized service recommendation device with the same or similar functions. This embodiment and the following embodiments will be described using a personalized service recommendation device as an example.

[0077] Based on this, the embodiments of this application provide a personalized service recommendation method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the personalized service recommendation method of this application.

[0078] In this embodiment, the personalized service recommendation method includes steps S10 to S90:

[0079] Step S10: Collect customer feedback data from various channels through a pre-set platform.

[0080] Understandably, multi-channel customer feedback collection systems typically utilize various channels such as telephone, email, and social media to gather customer feedback. Businesses use these channels to receive customer opinions, suggestions, and complaints, thereby acquiring valuable customer data. However, feedback data from each channel is relatively independent, lacking unified and comprehensive analysis. Therefore, this embodiment uses a unified platform to collect customer feedback from different channels (such as telephone, email, and social media). The data collection formula comprehensively considers channel weights and feedback influencing factors to ensure the representativeness and validity of the feedback data.

[0081] Step S20: Perform feedback sentiment analysis based on the customer feedback data to obtain the feedback sentiment analysis results.

[0082] It should be understood that feedback sentiment analysis can be performed based on customer feedback data. Each piece of feedback data from each customer is input into a sentiment classification model, and a sentiment score for each piece of feedback data from each customer is output. Then, the sentiment scores are weighted to obtain the feedback sentiment analysis results.

[0083] In one feasible implementation, step S20 may include steps S201 to S204:

[0084] Step S201: Determine the feedback weight of each customer's feedback data in the customer feedback data based on the feedback importance weight, time-related weight, and feedback source weight.

[0085] It is understandable that η is used ij Let represent the feedback weight of the j-th feedback from the i-th customer, calculated using the following formula:

[0086] η ij =η impij ×η timeij ×η srcij ,

[0087] Where, η impij η represents the importance weight of the feedback. timeij For time-related weights, η srcij The weight of the feedback source.

[0088] Specifically, the importance weight η of the feedback impij for:

[0089]

[0090] Among them, importance ij The importance of the j-th feedback from the i-th customer;

[0091] Time correlation weight η timeij for:

[0092]

[0093] Where e is the natural constant, ∈ is the decay coefficient, and t is the current time. ij The feedback time is the j-th feedback from the i-th customer;

[0094] Importance of feedback sources η srcij for:

[0095]

[0096] Wherein, source ij Rate the source of the j-th feedback from the i-th customer.

[0097] Step S202: Based on the feedback weight, the emotional score of each customer's feedback data in the customer feedback data is weighted and summed to obtain the feedback weight summation result.

[0098] It is understandable that s is used ij Let s represent the sentiment score of the j-th feedback from the i-th customer. ij The set of values ​​is {-1, 1}, where -1 represents negative feedback and 1 represents positive feedback. The sentiment score s of the j-th feedback from the i-th customer is... ij The sentiment score is calculated using a sentiment classification model. Specifically, the sentiment score s is calculated by inputting the j-th feedback from the i-th customer into the sentiment classification model, and the model outputs the sentiment score s of the j-th feedback from the i-th customer. ij This sentiment classification model can be trained based on a predefined sentiment lexicon, which lists words with positive or negative sentiment tendencies, and each word has a corresponding sentiment score (e.g., -1 for negative sentiment, +1 for positive sentiment). The aforementioned sentiment classification model includes, but is not limited to, support vector machines, random forests, and deep neural networks.

[0099] It should be understood that, based on the feedback weights, the sentiment scores of each customer's feedback data are weighted and summed to obtain the feedback weight summation result:

[0100]

[0101] Step S203: Determine the customer points reward value for each customer based on the customer emotional influence factor and customer points reward.

[0102] It is understandable that β is used. i R represents the customer sentiment influencing factor. i This represents the customer points reward for the i-th customer. The customer points reward value for each customer is calculated as the product of the customer sentiment influence factor and the customer points reward: β i R i .

[0103] Step S204: Determine the feedback sentiment analysis result for each customer in the customer feedback data based on the summation result of the feedback weights and the customer points reward value.

[0104] It should be noted that, based on the summation of the feedback weights and the customer points reward value, the sentiment analysis result for each customer in the customer feedback data is determined as follows:

[0105]

[0106] Among them, S iThe result of the sentiment analysis is used to provide feedback, which represents the weighted overall sentiment score of the i-th customer.

[0107] In this embodiment, the sentiment score of each customer's feedback data is output through the sentiment classification model. The feedback weight is determined according to the weight of the feedback importance, the weight of the time relevance, and the weight of the feedback source. The sentiment scores are weighted and summed according to the feedback weight. Then, the customer sentiment influence factor and the customer points reward are introduced to conduct feedback sentiment analysis, so as to more accurately calculate the feedback sentiment analysis results of each customer's feedback data.

[0108] Step S30: Calculate customer satisfaction score based on the customer feedback data to obtain a customer satisfaction score value.

[0109] Understandably, customer satisfaction data can be collected through questionnaires, rating systems, or direct feedback. Then, customer satisfaction scores can be calculated based on the satisfaction data. The customer satisfaction score value is obtained by combining the satisfaction score of each customer for each rating with the rating weight and influencing factors.

[0110] In one feasible implementation, step S30 may include steps S301 to S303:

[0111] Step S301: Based on the satisfaction rating weights, the satisfaction rating scores of each customer in the customer feedback data are weighted and summed to obtain the satisfaction rating summation result.

[0112] It should be noted that when using m ij η represents the satisfaction score of the i-th customer in the j-th rating. ij This represents the satisfaction rating weight of the i-th customer in the j-th rating. The two weighted sums are then used to obtain the satisfaction rating sum:

[0113]

[0114] The formula for calculating the weight of the satisfaction score is as follows:

[0115] η ij =η impij ×η timeij ×η srcij ,

[0116] Where, η impij η represents the importance weight of the feedback. timeij For time-related weights, η srcij The weight of the feedback source.

[0117] Specifically, the importance weight η of the feedback impij for:

[0118]

[0119] Among them, importance ij The importance of the j-th feedback from the i-th customer;

[0120] Time correlation weight η timeij :

[0121]

[0122] Where e is the natural constant, ∈ is the decay coefficient, and t is the current time. ij The feedback time is the j-th feedback from the i-th customer;

[0123] Importance of feedback sources η srcij :

[0124]

[0125] Wherein, source ij Rate the source of the j-th feedback from the i-th customer.

[0126] Step S302: Determine the satisfaction impact value of each customer based on the satisfaction impact factor and the feedback sentiment analysis results of each customer.

[0127] It is understandable that δ is used i S represents the factor influencing customer sentiment on satisfaction scores. i Let represent the weighted overall sentiment score of the i-th customer. The satisfaction impact value for each customer, calculated by multiplying the two scores, is represented as: δ i S i .

[0128] Step S303: Determine the customer satisfaction score for each customer in the customer feedback data based on the summation result of the satisfaction and the influence value of the satisfaction.

[0129] Understandably, the customer satisfaction score for each customer in the customer feedback data is determined based on the sum of satisfaction scores and the satisfaction impact value:

[0130]

[0131] Where Mi is the customer satisfaction score, representing the weighted satisfaction score of the i-th customer.

[0132] Step S40: Identify potential problems based on the feedback sentiment analysis results and the customer satisfaction score to obtain potential problem identification results.

[0133] Understandably, the existence of potential problems can be determined based on pre-set thresholds, feedback sentiment analysis results, and customer satisfaction scores, resulting in potential problem identification results, which include both presence and absence of potential problems.

[0134] In one feasible implementation, step S40 may include steps S401 to S405:

[0135] Step S401: Adjust the historical feedback emotion score and historical customer satisfaction score data according to the adjustment coefficient to obtain the dynamic adjustment factor.

[0136] It should be noted that λ is used. i This represents a dynamic moderating factor for customer feedback sentiment and satisfaction. The dynamic moderating factor can be determined based on historical data analysis. Customer feedback sentiment scores (S) are collected over a period of time. i and satisfaction score M i ,

[0137]

[0138] Among them, D ij =|S ij -M ij | represents the difference between the emotion score and satisfaction score of the i-th customer in the j-th instance. Let N be the mean of all differences for the i-th customer, and N be the total number of data points. The dynamic adjustment factor can be set as a multiple of the standard deviation to adjust sensitivity based on the fluctuation range of historical data.

[0139] λ i =k i ×σ i

[0140] Where, k i This is an adjustment coefficient, which can be adjusted according to the actual situation. i ∈(0,1).

[0141] Step S402, or, set the adjustment coefficient according to the business objectives, and determine the dynamic adjustment factor according to the adjustment coefficient.

[0142] Understandably, dynamic adjustment factors can also be set based on experience with business objectives. This involves determining the company's required sensitivity to customer feedback, such as wanting to identify potential problems more promptly or allowing for more lenient feedback adjustments. The adjustment coefficient k is then set according to these business objectives. i The adjustment is made by comparing the effects under different adjustment coefficients. The set adjustment coefficients are verified through practical application, and adjustments are made based on feedback to ensure that the adjustment factor can effectively identify potential problems.

[0143] Step S403, or, using machine learning to predict historical feedback sentiment scores and historical customer satisfaction scores, and determining dynamic adjustment factors based on the prediction results.

[0144] It should be understood that dynamic adjustment factors can also be determined using machine learning-based methods. This can be achieved by collecting a large amount of customer feedback sentiment scores (S). ij and satisfaction score M ij and its differences D ij =|S ij -M ij Train a regression model to predict the difference between customer feedback sentiment and satisfaction, and adjust the dynamic adjustment factor based on the prediction results. Using the trained model, calculate the dynamic adjustment factor in real time based on the current sentiment and satisfaction scores.

[0145] Step S404: Determine the difference between the feedback sentiment analysis result and the customer satisfaction score.

[0146] Step S405: Determine the potential problem identification result based on the difference, the preset difference threshold, and the dynamic adjustment factor.

[0147] It should be noted that potential problems can be identified using the following formula:

[0148]

[0149] Among them, P i To determine whether there are potential problems for the i-th customer, σ i λ is the threshold for the difference between the emotion score and the satisfaction score. i This is a dynamic adjustment factor for customer feedback on emotion and satisfaction. A value of 0 indicates that no potential problem exists, while a value of 1 indicates that a potential problem exists.

[0150] Step S50: Determine the customer lifecycle stage based on the feedback sentiment analysis results.

[0151] Understandably, in order to develop more precise personalized solutions, it is possible to combine the customer lifecycle stage with the solution. The customer lifecycle stage can be divided into multiple stages, such as introduction, growth, maturity and decline.

[0152] In one feasible implementation, step S50 may include steps S501 to S503:

[0153] Step S501: Determine the linear correlation between historical feedback sentiment scores and historical life cycle stage data based on the Pearson correlation coefficient, and adjust the linear correlation based on the adjustment coefficient to obtain the life cycle adjustment factor.

[0154] It should be noted that θ can be used to represent the moderating factor of sentiment score on lifecycle. The moderating factor θ can be determined in two ways: lifecycle factor identification based on statistical analysis or lifecycle factor identification based on business experience.

[0155] Specifically, the lifecycle factor identification method based on statistical analysis includes: collecting sufficient customer sentiment scores (S). i and life cycle stage L i Data; using the Pearson correlation coefficient to measure sentiment score S i and life cycle stage L i Linear correlation between them:

[0156]

[0157] Where N is the number of customers. The average of the sentiment scores. This is the average value over the lifespan;

[0158] θ = k × r,

[0159] Where k is an adjustment coefficient, which can be adjusted according to the needs of the enterprise;

[0160] Lifecycle factor identification based on business experience includes: determining the degree of influence of sentiment scores on emotional scores in each lifecycle stage based on the company's specific business objectives and industry standards, and setting appropriate adjustment factors θ through the opinions of business experts and market research.

[0161] Step S502: Determine the customer activity level of each customer based on online behavior data analysis, transaction data analysis, and interaction data analysis.

[0162] It is understandable that using a i Let a represent the activity level of the i-th customer. iThis refers to one or a comprehensive score based on online behavioral data analysis, transaction data analysis, and interaction data analysis. Online behavioral data analysis includes calculating activity levels based on customer visit frequency, dwell time, and clicks on websites or applications. For example, customer activity can be measured by the number of visits, purchases, and comments. Transaction data analysis includes calculating activity levels based on customer transaction frequency and amount. For example, customer activity can be measured by the number of transactions and transaction amount. Interaction data analysis includes calculating activity levels based on the frequency and depth of customer interactions with the business, such as phone calls, emails, and social media interactions. For example, a customer service center can measure customer activity based on the number of calls, email interactions, and social media interactions. Comprehensive scoring methods include combining multiple data sources and calculating a comprehensive customer activity score using weighted averaging or machine learning models. For example, considering visit frequency, transaction frequency, and interaction frequency, a weighted average can be used to calculate the comprehensive customer activity score.

[0163] Step S503: Determine the customer lifecycle stage based on the customer activity level, the feedback sentiment analysis result, the first activity threshold, the second activity threshold, the first feedback sentiment score threshold, the second feedback sentiment score threshold, and the lifecycle adjustment factor. The customer lifecycle stage includes the introduction stage, the growth stage, the maturity stage, and the decline stage. The first feedback sentiment score threshold is greater than or equal to the second feedback sentiment score threshold.

[0164] It's worth noting that, based on customer data, the customer lifecycle is divided into four stages: introduction, growth, maturity, and decline. The specific division method is as follows:

[0165]

[0166] Among them, L i For the i-th customer's lifecycle stage, a i For the activity level of the i-th customer, a th1 a th2 The threshold representing activity, S i S is the weighted overall sentiment score for the i-th customer. th1 and S th2 The threshold representing the sentiment score, typically S th1 >S th2 Of course, S can also be chosen. th1 =S th2 θ is the moderating factor of emotional score on life cycle.

[0167] Understandably, during the introduction phase, when customers are just beginning to engage with the product or service, customer activity is low. iTypically, the level is low, and the customer's emotional score may not have fully formed. Therefore, customer activity level is used as the primary criterion, with customer emotional score as a secondary criterion, to ultimately determine whether the product is in the introduction phase. Similarly, in the decline phase, customer activity level a... i Customer sentiment scores are typically low, often negative. Therefore, customer activity is used as the primary criterion, with customer sentiment score serving as a secondary criterion to determine whether a business is in a decline phase. During the growth and maturity phases, the relationship between customer sentiment score and activity becomes more complex, at which point a moderating factor θ is introduced.

[0168] Step S60: Determine the points reward based on the customer lifecycle stage and the customer satisfaction score.

[0169] It should be noted that the formula for calculating points rewards is as follows:

[0170] R i =α1C i +α2A i +α3M i +α4L i ,

[0171] Among them, R i For the points reward of the i-th customer, α1 is the weighting coefficient, and C i For customer spending, A i For customer activity, α2 is the weighting coefficient, and M is the customer activity level. i Let α3 be the satisfaction score for the i-th customer, and L be the weighting coefficient. i Let α4 be the lifecycle of the i-th customer, and α4 be the weighting coefficient.

[0172] Step S70: Determine exclusive offers based on the customer lifecycle stage and the points reward.

[0173] Understandably, the formula for calculating exclusive discounts is:

[0174] D i =f(M i L i P i R i S i )

[0175] Among them, D i For the discount for the i-th customer, M i For customer satisfaction, L i For customer lifecycle stages, P i For potential customer issues, R i S is a points reward program for customers. i Assess the customer's emotional state.

[0176] Specifically, D i =β 11 M i +β 12 L i +β 13 P i +β 14 R i +β 15 S i +β 16 (M i ×L i )+C,

[0177] Where, β 11 β 12 β 13 β 14 β 15 β 16 M is a linear weighting coefficient, and C is a constant offset used to adjust the base discount value. i ×L i This is an interaction between customer satisfaction and customer lifecycle. On the one hand, it is used to differentiate the different sensitivities of customers to satisfaction at different stages of the customer lifecycle. For example, customers in the introduction or growth stage may be more sensitive to changes in service experience and satisfaction. On the other hand, it helps insurance platforms adjust their preferential strategies according to the customer's lifecycle stage. For example, for mature customers, even if their satisfaction is relatively low, higher discounts may be offered to ensure their loyalty, while for introduction customers, high satisfaction may require greater discounts to cultivate their long-term value.

[0178] or,

[0179]

[0180] Wherein, log(L) i The "2" in +2) is used to avoid inputting zero. The "2" can be adjusted according to the actual situation. (1-P) i This is used to ensure that the impact of potential problems is also within a reasonable range. and Able to avoid R i and S i The adverse effects of excessive β 21 β 22 β 23 β 24 β 25 β 26 These are coefficients determined based on actual circumstances.

[0181] Step S80: Based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward, a comprehensive analysis is performed to obtain comprehensive behavioral characteristics.

[0182] It should be understood that, based on a comprehensive analysis of the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward, the formula for calculating the comprehensive behavioral characteristics is as follows:

[0183] B i =f(L i M i S i a i P i R i ),

[0184] Among them, B i L is the result of a comprehensive analysis of the behavioral characteristics of the i-th customer. i For the i-th customer's lifecycle stage, M i For the satisfaction rating of the i-th customer, S i Let a be the weighted overall sentiment score for the i-th customer. i For the activity level of the i-th customer, P i To determine whether there are potential problems for the i-th customer, R i The points reward for the i-th customer.

[0185] Specifically, B i =γ1L i +γ2M i +γ3S i +γ4a i +γ5P i +γ6R i Among them, γ1, γ2, γ3, γ4, γ5 and γ6 are weighting coefficients, which are obtained based on practical experience.

[0186] Alternatively, a comprehensive analysis can be performed using a random forest regression model: B i =RandomForestRegressor(L i M i S i a i P i R i ).

[0187] Step S90: Determine a personalized service recommendation plan based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, and push the personalized service recommendation plan to the user.

[0188] It is understandable that the interaction flow of the personalized service recommendation method in this application can be referenced. Figure 2 Personalized service recommendations are determined based on customer feedback data, sentiment analysis results, customer satisfaction scores, potential issues, customer lifecycle stages, points rewards, exclusive offers, and overall behavioral characteristics. The calculation formula is as follows:

[0189] PS i =g(F i L i S i M i P i B i R i D i )

[0190] Among them, PS i For the personalized service recommendation of the i-th customer, F i For the feedback data of the i-th customer, L i For the i-th customer's lifecycle stage, S i For the emotional score of the i-th customer, M i For the satisfaction score of the i-th customer, P i For the potential problems of the i-th customer, B i R is the result of a comprehensive analysis of the behavioral characteristics of the i-th customer. i D represents customer points rewards. i Exclusive offers for our clients. A comprehensive analysis can be performed using a random forest regression model.

[0191] PS i =RandomForestRegressor(F i L i S i M i P i B i R i D i ).

[0192] This embodiment provides a personalized service recommendation method. It collects customer feedback data from various channels through a pre-set platform. Based on this data, it determines feedback sentiment analysis results, customer satisfaction scores, potential problems, customer lifecycle stages, points rewards, exclusive offers, and comprehensive behavioral characteristics, and then generates a personalized service recommendation scheme. This application ensures the representativeness and validity of the feedback data by collecting it through a unified platform. Real-time analysis of the collected feedback data, focusing on feedback sentiment, customer satisfaction, potential problem identification, lifecycle stages, points rewards, and exclusive offers, accurately identifies customers' true emotions and needs, ensuring the accuracy and reliability of the analysis results. Through comprehensive behavioral characteristic analysis, it provides personalized service recommendations, thereby improving customer satisfaction.

[0193] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 Step S10, the personalized service recommendation method further includes steps S101 to S106:

[0194] Step S101: Collect customer feedback data from various channels through a pre-set platform.

[0195] It should be understood that in this embodiment, data is collected uniformly through a pre-set platform to ensure the representativeness and validity of the feedback data. Customer feedback is collected from different channels (such as telephone, email, and social media): using F... t Let represent the total feedback amount (customer feedback data) at time t, and fi(t) represent the feedback data of the i-th channel at time t.

[0196] It should be noted that feedback data can include multi-dimensional data such as the claims process experience, customer service experience, policy purchase and renewal experience, understanding and satisfaction with insurance products, emergency support and services, the gap between experience and expectations, insurance promotion and marketing activities, follow-up services and customer care, and the reasons and explanations for insurance claim rejections.

[0197] Specifically, the claims process experience includes the simplicity and transparency of the claims application process, whether the claims processing time meets expectations, whether the claim amount is reasonable and fair, and the smoothness of communication with the insurance company during the claims process. Customer service experience includes the professionalism and attitude of customer service personnel, the timeliness of customer service response, the clarity and accuracy of customer service in explaining insurance terms and benefits, and whether problems or complaints are effectively resolved. Policy purchase and renewal experience includes the convenience of the policy purchase process, the choice and flexibility of premium payment methods, the ease of understanding policy terms, and the smoothness of the policy renewal or renewal process. Insurance product understanding and satisfaction includes overall satisfaction with insurance products and services, satisfaction with insurance coverage, perception of the cost-effectiveness of premiums and benefits, and whether the insurance terms are clear and easy to understand. Emergency support and services include the speed and level of support from the insurance company in emergency situations, feedback on additional services provided by the insurance company (such as roadside assistance, health consultation, etc.), the effectiveness of communication and problem-solving efficiency in emergency situations. The gap between experience and expectations includes whether insurance services meet expectations, feedback on unexpected situations in the actual use of benefits, and the insurance company's attitude and effectiveness in handling emergencies. Insurance promotion and marketing activities include opinions on the insurance company's promotional activities or discounts, and feedback on the completeness and transparency of the information provided in the promotional activities. Follow-up services and customer care include whether the insurance company provides continuous customer care and services during the policy's effective period, whether the customer receives regular policy information updates and usage reminders, and the quality of follow-up services after the use of benefits. Reasons and explanations for insurance claim denials include whether the explanation for the claim denial is clear and reasonable, the level of understanding and acceptance of the reasons for the denial, and satisfaction with the appeal or reprocessing channels after the claim denial.

[0198] Step S102: Determine the channel weights corresponding to the various channels based on customer coverage, feedback quality, customer preferences, and response speed.

[0199] It is understandable that ω is used. i Channel weight can be calculated using the following formula:

[0200] ω i =α co ω covi +β qu ω quali +γ pr ω prei +δ re ω respi ,

[0201] Where, α co β qu γ pr and δ reThese are weighting coefficients. Channel weighting can consider the following four aspects: Channel customer coverage: Some channels may cover more customers, therefore their feedback may be more representative. Channel feedback quality: Some channels provide higher quality, more specific, and more useful feedback. Channel customer preference: Customers prefer to provide feedback through certain channels. Channel response speed: Some channels provide more real-time feedback, which helps to adjust and improve services promptly.

[0202] Specifically, channel weights from different perspectives can be calculated by using a weighted average. Among these, the channel's customer coverage rate weight ω... covi The calculation is as follows:

[0203]

[0204] Among them, COV i Let n be the customer coverage rate of the i-th channel, and n be the total number of channels. This represents the total customer coverage across all channels, used to normalize coverage weights.

[0205] Channel feedback quality weight ω quali The calculation is as follows:

[0206]

[0207] Among them, QUAL i Let n represent the feedback quality of the i-th channel, and n be the total number of channels. This is the sum of feedback quality from all channels, used to normalize the feedback quality weights.

[0208] Channel customer preference weight ω prei :

[0209]

[0210] Where PREi represents the customer preference level for the i-th channel, and n is the total number of channels. This is the sum of customer preferences across all channels, used to normalize customer preference weights.

[0211] Channel response speed weight ω respi :

[0212]

[0213] Among them, V i Let be the response speed of the i-th channel, and n be the total number of channels. This is the sum of response speeds across all channels, used to normalize response speed weights.

[0214] Step S103: The feedback data from the various channels are weighted and summed according to the channel weights to obtain the channel weight summation result.

[0215] It is understandable that the feedback data from the various channels can be weighted and summed according to the channel weights to obtain the channel weight summation result as follows:

[0216]

[0217] Step S104: Determine the feedback influencing factors based on feedback timeliness, customer lifecycle stage, market environment changes, and historical feedback data.

[0218] Understandably, αt is used to represent the feedback influence factor at time t, indicating the degree of importance that specific environmental, market changes, or corporate strategic adjustments place on feedback at the current moment. For example, when the market environment changes drastically or a new product is launched, a company may need a higher feedback weight to respond quickly to customer needs.

[0219] Understandably, the feedback impact factor can be calculated as follows:

[0220] α t =α ti ω timei +β li ω lifei +γ en ω envi +δ hi ω histi

[0221] In other words, the feedback impact factor αt needs to consider the following four aspects: Timeliness of feedback: Feedback at different points in time has varying importance; real-time feedback is more valuable. Customer lifecycle stage: Feedback from customers at different lifecycle stages has different impacts on corporate decisions. Changes in the market environment: Changes in the external market environment also affect the importance of feedback. Historical feedback data: Trends in historical feedback data can help adjust the impact factor.

[0222] Specifically, in the formula for calculating the feedback impact factor mentioned above, α ti This is a coefficient related to the timeliness of feedback. This coefficient is used to adjust the weight of feedback timeliness in the overall feedback impact factor. Timely feedback is generally more valuable, therefore its influence needs to be appropriately amplified. β li This coefficient is related to customer lifecycle stages. Different customer lifecycle stages have different impacts on insurance platforms; for example, feedback from new customers may be more indicative of product or service issues than feedback from established customers. This coefficient adjusts the weight of lifecycle stage in the feedback influencing factors and is relevant to the specific decisions made by the insurance platform. γ enThis coefficient, δ, reflects the moderating effect of environmental factors on feedback, as changes in the market environment (such as increased competition or economic fluctuations) influence the importance and interpretation of feedback. hi This is a coefficient related to historical data. This coefficient is used to adjust the impact of historical data trends on current feedback. By observing the changing trends of historical data, the weight of current feedback can be better predicted and adjusted.

[0223] The weight ω of the timeliness of feedback timei :

[0224]

[0225] Among them, TIME i Let represent the timeliness of feedback at time i, indicating the time taken from information collection to decision-making, where n is the total number of collection points. The sum of timeliness across all time points is used to normalize the weights of feedback timeliness.

[0226] Customer lifecycle stage weight ω lifei :

[0227]

[0228] Among them, LIFE i To determine the importance of a customer at time i in the lifecycle stage, where n is the number of lifecycle stages, in this technical solution n = 4. It represents the sum of importance across the entire lifecycle and is used to normalize the weights of different stages in the customer lifecycle.

[0229] Weight ω of changes in the market environment envi :

[0230]

[0231] Among them, ENV i Let n represent the importance of the market environment at time i, and n be the total number of observation times of the market environment. Weights used to normalize market changes.

[0232] The weight ω of historical data histi :

[0233]

[0234] Among them, HIST i Let n represent the trend of historical feedback data at time i, and n be the total number of historical feedback data collection points. Weights used for normalizing historical data.

[0235] Step S105: Determine the feedback impact value based on the feedback impact factor and the life cycle stage corresponding to the previous time of the preset time.

[0236] It is understandable that L is used. i t-1 represents the lifecycle stage at time t-1, indicating the customer's importance to the insurance platform at that stage. Feedback from customers at different lifecycle stages may have varying importance for the company's decisions. For example, feedback from customers in the growth stage may be more helpful in optimizing products, while feedback from customers in the decline stage may reflect more problems with the product or service. The feedback impact factor is obtained by multiplying it by the lifecycle stage corresponding to the previous time point (previous time). t L i，t-1 This feedback impact value gives higher weight to feedback from customers at important lifecycle stages (such as growth or maturity) at key moments (such as when new policies or services are launched). This design ensures that the insurance platform pays more attention to feedback from these key customers when a rapid response is needed. Specifically, when the feedback impact factor α... t When the value is large, it indicates that the current moment is a feedback moment that the insurance platform particularly values; in this case, multiplying by a larger L... i，t-1 (For example, customers in their growth stage) will have their feedback weighted more significantly, prompting insurance platforms to prioritize and analyze this feedback; when the feedback impact factor α... t When it is small, even if L i,t-1 Larger, α t L i,t-1 The feedback weight will also be smaller, thus reducing the waste of resources.

[0237] Step S106: Sum the channel weight summation result and the feedback impact value to obtain customer feedback data.

[0238] It is understandable that the summation of the channel weights and the feedback impact value are summed to obtain the customer feedback data F. t , can be represented as follows:

[0239]

[0240] In this embodiment, customer feedback from different channels is collected through a unified platform. The feedback data from various channels is weighted and summed to obtain the channel weight summation result. The feedback impact value is determined by combining the feedback impact factor and the lifecycle stage corresponding to the previous time. The channel weight summation result and the feedback impact value are summed to obtain more effective and accurate customer feedback data.

[0241] Based on the first embodiment of this application, in the third embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 4 After step S90, the personalized service recommendation method further includes steps A10 to A50:

[0242] Step A10: Evaluate the personalized service recommendation scheme and obtain the evaluation results.

[0243] It should be understood that this embodiment ensures the continuous improvement of customer satisfaction and loyalty by evaluating and optimizing the effectiveness of personalized service recommendations in real time. Customer satisfaction and loyalty can be determined based on feedback sentiment analysis, customer satisfaction scores, points rewards, and exclusive offers.

[0244] U = h(M, R, D, S),

[0245] Where U represents the overall improvement in customer satisfaction and loyalty, M represents average satisfaction, R represents points rewards, D represents exclusive offers, and S represents emotional score. Specifically,

[0246]

[0247] Where, μ 11 μ 12 μ 13 and μ 14 As a coefficient, it allows for flexible control over the influence of each factor;

[0248] or,

[0249]

[0250] Where, μ 21 μ 22 μ 23 μ 24 μ is a coefficient used to adjust for the nonlinear effects of different factors on U. 25 As an overall bias, it can prevent the output value from being too large or too small.

[0251] Personalized service recommendations can be optimized based on customer satisfaction and loyalty. The personalized service recommendation scheme is then evaluated to obtain the evaluation result Ei:

[0252] ΔM i =M i，after -M i，before ,

[0253] ΔU i =U i，after -U i，before ,

[0254] E i =ΔM i +ΔU i ,

[0255] Among them, E i The effectiveness of personalized service recommendations for the i-th customer.

[0256] Step A20: Determine whether the personalized service recommendation scheme is effective based on the evaluation results.

[0257] It is understandable that if E i If E > 0, then personalized service recommendations are effective; if E i If the value is ≤0, then the personalized service recommendation needs to be adjusted.

[0258] Step A30: If the personalized service recommendation scheme is invalid, then adjust the customer's potential problems according to the scheme adjustment coefficient and the evaluation results, and determine the adjusted personalized service recommendation scheme according to the adjusted customer potential problems.

[0259] It should be understood that if the personalized service recommendation scheme is invalid, it will be adjusted. The adjustment process for personalized service recommendations is as follows: Original recommendation scheme: PS i =g(F i L i S i M i P i B i R i D i The adjustment process is as follows: P i ′=P i ×(1-α P ×E i ), α P Adjustment factors obtained based on experience or historical data trials may be used to address potential problems that customers may encounter. i Zoom in or out. The recommended adjustment is: Photoshop i =g(F i L i S i M i P i ′, B i R i D i ).

[0260] Step A40: Adjust the points reward according to the adjustment weighting coefficient to obtain the adjusted points reward.

[0261] It should be noted that, to further optimize the effectiveness of personalized service recommendations, it is possible to base it on E i The points reward mechanism has been optimized; please refer to the adjusted diagram. Figure 5 The bolded section in the image shows the process for adjusting the points reward system. The points reward system has been adjusted as follows:

[0262] D i =R i =α1C i +α2A i +α3M i +α4L i -α D ×E i ,

[0263] Where, α D These are the weighting coefficients.

[0264] Step A50: Optimize the adjusted personalized service recommendation scheme according to the adjusted points reward, obtain the optimized personalized service recommendation scheme, and push the optimized personalized service recommendation scheme to the user.

[0265] Understandably, adjustments to points rewards will be further reflected in adjustments to exclusive offers and the confirmation process of comprehensive behavioral characteristics, ultimately impacting customer loyalty and personalized service recommendations. For customers, adjustments to points rewards will also affect their feedback; this is a hidden influence created by the customers themselves. Therefore, this adjustment based on E... i Adjust points reward D i The recommendation method can combine dynamic customer feedback in the market to achieve dynamic recommendations, thereby further improving customer satisfaction and loyalty.

[0266] Furthermore, after obtaining customer feedback data in step S10, clustering algorithms can be used to classify the customer feedback, identifying commonalities and differences. In subsequent feedback sentiment analysis and customer satisfaction scoring steps, analysis and scoring can be based on the commonalities in the clustering results, i.e.: S2, perform feedback sentiment analysis based on common customer feedback; S3, score customer satisfaction based on common customer feedback. This can reduce the adverse effects of discrete data (such as malicious feedback) on the feedback sentiment analysis results and satisfaction scoring results, and improve the accuracy of the feedback sentiment analysis results and satisfaction scoring results. The specific scheme for classifying customer feedback is as follows:

[0267] C = k-means(F, k, γS)

[0268] Where C represents the clustering result of the feedback, F is the data matrix of all customer feedback, k is the number of clusters, and γ is the weight of the sentiment score on the cluster. The clustering analysis process includes: collecting feedback data from all customers to form a data matrix F, where each row represents the feedback of one customer and each column represents a feedback feature (such as feedback score, feedback time, etc.); and assigning a weighted sentiment score S to each customer. i The feedback data matrix is ​​incorporated to form an expanded data matrix; the feedback data is standardized to ensure that the data have the same dimensions, which facilitates cluster analysis.

[0269] Specifically, suppose there are m customers and n feedback features, and the dimension of the feedback data matrix F is m×(n+1):

[0270]

[0271] Standardize each column of the feedback data matrix F′ so that its mean is 0 and its standard deviation is 1.

[0272]

[0273] Where, μ j Let σ be the mean of the column vectors. j The standard deviation of the column vector;

[0274]

[0275]

[0276] Choose the number of clusters k, and initialize the cluster centers μ1, μ2, ..., μ3. k ;

[0277] Calculate the Euclidean distance between each data point and each cluster center:

[0278]

[0279] Where, μ jS Let be the mean sentiment score of the j-th cluster;

[0280] Assign each data point to the nearest cluster center:

[0281] cluster(i) = arg min j d ij ,

[0282] Recalculate the center of each cluster.

[0283]

[0284] Among them, C jLet |C| be the set of data points in the j-th cluster. j | represents the number of data points in the j-th cluster; until the clustering results converge;

[0285] C = k-means(F′) std ,k,γS),

[0286] Where C represents the clustering result, γ represents the weight of the sentiment score on the clustering, and S represents the sentiment score.

[0287] In this embodiment, the personalized service recommendation scheme is evaluated. If the personalized service recommendation scheme is invalid, the potential customer problems are adjusted according to the scheme adjustment coefficient and the evaluation results, and the adjusted personalized service recommendation scheme is determined based on the adjusted potential customer problems. The points reward is adjusted according to the adjustment weight coefficient to obtain the adjusted points reward. The adjusted personalized service recommendation scheme is optimized according to the adjusted points reward, so that personalized service recommendations can be optimized in real time based on customer satisfaction and loyalty to obtain an optimized personalized service recommendation scheme.

[0288] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the personalized service recommendation method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0289] This application also provides a personalized service recommendation device, please refer to... Figure 6 The personalized service recommendation device includes:

[0290] The customer feedback data collection module 10 is used to collect customer feedback data from various channels through a preset platform.

[0291] The feedback sentiment analysis module 20 is used to perform feedback sentiment analysis based on the customer feedback data and obtain feedback sentiment analysis results.

[0292] The customer satisfaction rating module 30 is used to score customer satisfaction based on the customer feedback data and obtain a customer satisfaction rating value.

[0293] The potential problem identification module 40 is used to identify potential problems based on the feedback sentiment analysis results and the customer satisfaction score, and obtain potential problem identification results.

[0294] Customer lifecycle determination module 50 is used to determine the customer lifecycle stage based on the feedback sentiment analysis results;

[0295] The points reward determination module 60 is used to determine points rewards based on the customer lifecycle stage and the customer satisfaction score.

[0296] The exclusive offer determination module 70 is used to determine exclusive offers based on the customer lifecycle stage and the points rewards.

[0297] The behavioral feature comprehensive analysis module 80 is used to perform comprehensive analysis based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer life cycle stage, and the points reward to obtain comprehensive behavioral features;

[0298] The personalized service recommendation module 90 is used to determine a personalized service recommendation scheme based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, and to push the personalized service recommendation scheme to the user.

[0299] The personalized service recommendation device provided in this application, employing the personalized service recommendation method in the above embodiments, can solve the technical problem. Compared with the prior art, the beneficial effects of the personalized service recommendation device provided in this application are the same as those of the personalized service recommendation method provided in the above embodiments, and other technical features in the personalized service recommendation device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0300] This application provides a personalized service recommendation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the personalized service recommendation method in Embodiment 1 above.

[0301] The following is for reference. Figure 7 The diagram illustrates a structural schematic of a personalized service recommendation device suitable for implementing embodiments of this application. The personalized service recommendation device in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The personalized service recommendation device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0302] like Figure 7As shown, the personalized service recommendation device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the personalized service recommendation device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the personalization service recommendation device to communicate wirelessly or wiredly with other devices to exchange data. While the figure shows personalization service recommendation devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0303] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0304] The personalized service recommendation device provided in this application, employing the personalized service recommendation method in the above embodiments, can solve the technical problem of personalized service recommendation. Compared with the prior art, the beneficial effects of the personalized service recommendation device provided in this application are the same as those of the personalized service recommendation method provided in the above embodiments, and other technical features in this personalized service recommendation device are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0305] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0306] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0307] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the personalized service recommendation method in the above embodiments.

[0308] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0309] The aforementioned computer-readable storage medium may be included in the personalized service recommendation device; or it may exist independently and not be assembled into the personalized service recommendation device.

[0310] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0311] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0312] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0313] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described personalized service recommendation method, thereby solving the technical problem. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the personalized service recommendation method provided in the above embodiments, and will not be repeated here.

[0314] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the personalized service recommendation method described above.

[0315] The computer program product provided in this application can solve the technical problem. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the personalized service recommendation method provided in the above embodiments, and will not be repeated here.

[0316] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A personalized service recommendation method, characterized in that, The method includes: Customer feedback data from various channels is collected uniformly through a pre-set platform; Based on the customer feedback data, feedback sentiment analysis is performed to obtain feedback sentiment analysis results; Based on the customer feedback data, a customer satisfaction score is obtained. Based on the feedback sentiment analysis results and the customer satisfaction score, potential problems are identified to obtain potential problem identification results. The customer lifecycle stage is determined based on the feedback sentiment analysis results; The points-based reward is determined based on the customer lifecycle stage and the customer satisfaction score. Exclusive offers are determined based on the customer lifecycle stage and the points rewards; A comprehensive analysis is conducted based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward to obtain comprehensive behavioral characteristics; Based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, a personalized service recommendation plan is determined and pushed to the user.

2. The method as described in claim 1, characterized in that, The step of uniformly collecting customer feedback data from multiple different channels through a pre-set platform includes: Customer feedback data from various channels is collected uniformly through a pre-set platform; The channel weights for the various channels are determined based on customer coverage, feedback quality, customer preferences, and response speed. The feedback data from the various channels are weighted and summed according to the channel weights to obtain the channel weight summation result. The feedback influencing factors are determined based on feedback timeliness, customer lifecycle stage, changes in the market environment, and historical feedback data. The feedback impact value is determined based on the feedback impact factor and the life cycle stage corresponding to the previous time of the preset time; The summation of the channel weights and the feedback impact value are summed to obtain customer feedback data.

3. The method as described in claim 1, characterized in that, The step of performing feedback sentiment analysis based on the customer feedback data to obtain the feedback sentiment analysis results includes: The feedback weight of each customer's feedback data is determined based on the feedback importance weight, time relevance weight, and feedback source weight. Based on the feedback weights, the sentiment scores of each customer's feedback data in the customer feedback data are weighted and summed to obtain the feedback weight summation result; The customer points reward value for each customer is determined based on the customer sentiment influence factor and the customer points reward. The feedback sentiment analysis result for each customer in the customer feedback data is determined based on the summation result of the feedback weights and the customer points reward value.

4. The method as described in claim 1, characterized in that, The step of scoring customer satisfaction based on the customer feedback data to obtain a customer satisfaction score includes: The satisfaction scores of each customer in the customer feedback data are weighted and summed for each rating based on the satisfaction rating weights to obtain the satisfaction summation result. The satisfaction impact value of each customer is determined based on the satisfaction impact factors and the feedback sentiment analysis results of each customer. The customer satisfaction score for each customer in the customer feedback data is determined based on the summation of the satisfaction scores and the impact value of the satisfaction scores.

5. The method as described in claim 1, characterized in that, The step of identifying potential problems based on the feedback sentiment analysis results and the customer satisfaction score, and obtaining the potential problem identification results, includes: The dynamic adjustment factor is obtained by adjusting the historical feedback sentiment score and historical customer satisfaction score data according to the adjustment coefficient. Alternatively, adjustment coefficients can be set based on business objectives, and dynamic adjustment factors can be determined based on these adjustment coefficients. Alternatively, machine learning can be used to predict historical feedback sentiment scores and historical customer satisfaction scores, and dynamic adjustment factors can be determined based on the prediction results. Determine the difference between the feedback sentiment analysis result and the customer satisfaction score; The potential problem identification result is determined based on the difference, the preset difference threshold, and the dynamic adjustment factor.

6. The method as described in claim 1, characterized in that, The step of determining the customer lifecycle stage based on the feedback sentiment analysis results includes: The linear correlation between historical feedback sentiment scores and historical life cycle stage data was determined based on the Pearson correlation coefficient, and the linear correlation was adjusted based on the adjustment coefficient to obtain the life cycle adjustment factor. The customer activity level of each customer is determined based on online behavior data analysis, transaction data analysis, and interaction data analysis. The customer lifecycle stage is determined based on the customer activity level, the feedback sentiment analysis results, the first activity threshold, the second activity threshold, the first feedback sentiment score threshold, the second feedback sentiment score threshold, and the lifecycle adjustment factor. The customer lifecycle stage includes the introduction stage, the growth stage, the maturity stage, and the decline stage. The first feedback sentiment score threshold is greater than or equal to the second feedback sentiment score threshold.

7. The method as described in claim 1, characterized in that, After the step of determining a personalized service recommendation scheme based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, the method further includes: The personalized service recommendation scheme is evaluated to obtain the evaluation results; The effectiveness of the personalized service recommendation scheme will be determined based on the evaluation results. If the personalized service recommendation scheme is invalid, the customer's potential problems will be adjusted according to the scheme adjustment coefficient and the evaluation results, and the adjusted personalized service recommendation scheme will be determined according to the adjusted customer's potential problems. The points reward is adjusted according to the weighting coefficient to obtain the adjusted points reward; The adjusted personalized service recommendation scheme is optimized based on the adjusted points reward system to obtain an optimized personalized service recommendation scheme, which is then pushed to the user.

8. A personalized service recommendation device, characterized in that, The personalized service recommendation device includes: The customer feedback data collection module is used to collect customer feedback data from various channels through a pre-set platform. The feedback sentiment analysis module is used to perform feedback sentiment analysis based on the customer feedback data and obtain feedback sentiment analysis results. The customer satisfaction rating module is used to score customer satisfaction based on the customer feedback data and obtain a customer satisfaction rating value. The potential problem identification module is used to identify potential problems based on the feedback sentiment analysis results and the customer satisfaction score, and to obtain potential problem identification results. The customer lifecycle determination module is used to determine the customer lifecycle stage based on the feedback sentiment analysis results. The points reward determination module is used to determine points rewards based on the customer lifecycle stage and the customer satisfaction score. The exclusive offer determination module is used to determine exclusive offers based on the customer lifecycle stage and the points rewards. The behavioral characteristic comprehensive analysis module is used to perform comprehensive analysis based on the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, and the points reward to obtain comprehensive behavioral characteristics; The personalized service recommendation module is used to determine a personalized service recommendation plan based on the customer feedback data, the feedback sentiment analysis results, the customer satisfaction score, the potential problems, the customer lifecycle stage, the points rewards, the exclusive offers, and the comprehensive behavioral characteristics, and to push the personalized service recommendation plan to the user.

9. A personalized service recommendation device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the personalized service recommendation method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the personalized service recommendation method as described in any one of claims 1 to 7.

Images