Business processing method and device based on user portrait, equipment and medium

Abstract

The application relates to the technical field of intelligent decision-making, and discloses a business processing method and device based on user portraits, equipment and a medium, which comprises the following steps: receiving business feedback information, performing semantic analysis on the business feedback information, classifying the business feedback information according to the analysis result to obtain a user feedback intention; obtaining historical business behavior data, and constructing a user portrait according to the historical business behavior data; identifying a user type label of a target user, scaling and adjusting a preset initial label weight of the user type label according to an activity feature in the user portrait to obtain a label weight; adjusting the label weight according to a consumption feature and a credit feature to obtain a target user type feature value; when the target user type feature value exceeds a preset feature threshold value, calculating the matching degree of the user feedback intention and each business processing strategy, and screening a business processing strategy from a business strategy library. The application improves the accuracy of business processing.

Description

Technical Field

[0001] This invention relates to the field of intelligent decision-making technology, and in particular to a business processing method, apparatus, device, and medium based on user profiles. Background Technology

[0002] With the rapid development of internet technology, the number of users on various online service platforms (such as e-commerce, SaaS services, fintech, etc.) has surged. The number of business requests such as inquiries, complaints, and refunds submitted by users through feedback pages and customer service systems has also become massive and diverse.

[0003] Currently, traditional business processing methods employ simple strategies / keyword automation. For example, when the keyword "refund" is detected in user feedback, a standard message containing the refund process is automatically replied. However, this method relies solely on keyword matching, which can easily misjudge the user's true intent. It also fails to identify user value and risk, and cannot provide differentiated services to high-value or potentially churned users. This results in a high error rate in responding to users and reduces the accuracy of business processing.

[0004] Therefore, in the face of ever-increasing business processing demands, current business processing methods urgently need to be improved to address the problem of low accuracy in existing methods. Summary of the Invention

[0005] This invention provides a business processing method, apparatus, computer equipment, and medium based on user profiles to solve the problem of low accuracy in business processing.

[0006] Firstly, a business processing method based on user profiles is provided, including: Receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information according to the analysis results, and obtain the user feedback intent; Obtain historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data; construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. Identify the user type tags of the target user, and adjust the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights; The label weights are adjusted by attenuation based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights, and the attenuated label weights are used as the target user type feature values ​​of the target user. When the target user type feature value exceeds a preset feature threshold, the matching degree of the user feedback intent is calculated with each business processing strategy in the preset business strategy library used to process business feedback information. The business processing strategy with a matching degree greater than the preset matching threshold is selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategy.

[0007] Secondly, a business processing device based on user profiles is provided, including: The user feedback intent classification module is used to receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information into intents based on the analysis results, and obtain the user feedback intent. The user profile building module is used to acquire historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data, and build a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. The user type tag recognition module is used to identify the user type tag of the target user; The tag weight adjustment module is used to scale and adjust the preset initial tag weight of the user type tag according to the activity characteristics in the user profile to obtain the tag weight; The target user type feature value adjustment module is used to adjust the label weights by attenuation based on the consumption characteristics and credit characteristics in the user profile, to obtain the attenuated label weights, and to use the attenuated label weights as the target user type feature values ​​of the target user. The business feedback information processing module is used to calculate the matching degree between the user feedback intent and each business processing strategy in the preset business strategy library for processing business feedback information when the target user type feature value exceeds the preset feature threshold, filter out the business processing strategies with matching degree greater than the preset matching threshold from the business strategy library, and process the business feedback information according to the filtered business processing strategies.

[0008] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described user profile-based business processing method.

[0009] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described user profile-based business processing method.

[0010] The aforementioned solution, implemented using a user profile-based business processing method, apparatus, computer equipment, and storage medium, allows for semantic analysis of the business feedback information. Based on the analysis results, the feedback information is categorized by intent, accurately extracting the user's true needs and eliminating invalid and redundant information. Furthermore, by constructing a user profile based on consumption characteristics, credit characteristics, and activity characteristics, multi-dimensional information on consumption, credit, and activity is integrated to comprehensively depict user attributes and achieve refined user identification. Finally, by identifying the target user's user type tags and scaling the preset initial tag weights of the user type tags according to the activity characteristics in the user profile, the solution achieves a more refined user identification process. The baseline weights are adjusted based on the actual user activity level to avoid initial fixed weights deviating from actual user usage and improve the accuracy of business processing. The tag weights are attenuated and adjusted according to the consumption characteristics and credit characteristics, and the weights are adjusted by adding user value and risk dimensions to quantify user levels and improve the objectivity of user evaluation results. By calculating the matching degree between the user feedback intent and each business processing strategy in the preset business strategy library used to process business feedback information, the business processing strategies with a matching degree greater than a preset matching threshold are selected from the business strategy library, and the appropriate handling solutions are intelligently matched as needed to differentiate the handling of user feedback and improve the efficiency and rationality of feedback handling. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of an application environment for a user profile-based business processing method according to an embodiment of the present invention; Figure 2 This is a schematic flowchart of a business processing method based on user profiles in one embodiment of the present invention; Figure 3 This is a schematic diagram of a user profile-based business processing device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention; Figure 5 This is another structural schematic diagram of a computer device according to one embodiment of the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] The user profile-based business processing method provided in this invention can be applied to, for example... Figure 1 In this application environment, the client communicates with the server via a network. The server can receive business feedback information submitted by the target user to the business platform through the client, perform semantic analysis on the business feedback information, classify the intent of the business feedback information according to the analysis results, and obtain the user feedback intent; acquire historical business behavior data including the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data, construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics; identify the target user's user type tags, and assign preset initial tag weights to the user type tags according to the activity characteristics in the user profile. Scaling and adjustment are performed to obtain tag weights. These tag weights are then attenuated based on consumption and credit characteristics in the user profile, resulting in attenuated tag weights. These attenuated tag weights are used as the target user type feature value for the target user. When the target user type feature value exceeds a preset feature threshold, the user feedback intent is matched with each business processing strategy in a preset business strategy library for processing business feedback information. Business processing strategies with a matching degree greater than the preset matching threshold are selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategies, thereby improving the accuracy of business processing. The client can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented using a standalone server or a server cluster consisting of multiple servers. The invention will be described in detail below through specific embodiments.

[0015] Please see Figure 2 As shown, Figure 2 A flowchart illustrating a user profile-based business processing method provided in an embodiment of the present invention includes the following steps: S1. Receive business feedback information submitted by the target user to the business platform, perform semantic analysis on the business feedback information, classify the intent of the business feedback information according to the analysis results, and obtain the user feedback intent.

[0016] In this embodiment of the invention, the business feedback information refers to the text content submitted by the user through the front-end interface (such as a webpage or an in-app page) to express their needs, questions, opinions, or demands.

[0017] In this embodiment of the invention, the step of classifying the business feedback information according to the analysis results to obtain the user feedback intent includes: Based on the analysis results, the business feedback information is mapped to a preset set of intent categories with probability values; The intent category with the highest probability value in the intent category set is selected as the explicit intent, and the highest probability value is used as the confidence level of the explicit intent; Determine whether the confidence level is lower than a preset confidence threshold; When the confidence level is not lower than the confidence threshold, the explicit intent is taken as the user feedback intent. When the confidence level is lower than the confidence level threshold, the business feedback information is matched with the preset entity recognition strategy engine to obtain entity information; The attribute information of the entity is queried from a preset domain knowledge graph, and the attribute information includes common business issues; A candidate intent set is generated based on common issues in the attribute information. The similarity between the business feedback information and each candidate intent in the candidate intent set is calculated. The candidate intent with the highest similarity is selected as the implicit intent. The explicit intent and the implicit intent are combined into a user feedback intent.

[0018] In this embodiment of the invention, during the semantic analysis of business feedback information, a pre-trained natural language processing model is typically used. The business feedback information is taken as the input of the natural language processing model. The natural language processing model calculates the semantic vector of the business feedback information through a multi-layer neural network, and then matches the semantic vector with a predefined set of intent categories (such as "refund", "complaint", "inquiry", "exchange", etc.). A probability value is output for each intent category, representing the probability that the business feedback information belongs to that intent category, resulting in a structured set of intent category probability distributions, where each category is associated with a probability value between 0 and 1.

[0019] Furthermore, from the obtained set of probability distributions of intent categories, the intent category with the highest probability value is identified. For example, the probability of "refund" is 0.85, "inquiry" is 0.10, and "complaint" is 0.05. "Refund" is then selected as the explicit intent. At the same time, the highest probability value (0.85) is directly used as the confidence level of the explicit intent.

[0020] Furthermore, a confidence threshold is pre-set. If the confidence level is lower than the confidence threshold, the entity recognition strategy engine is invoked. This entity recognition strategy engine contains a series of rule-based or model-based entity extraction strategies (such as regular expression matching of order number, keyword matching of product name, named entity recognition model extraction of amount, etc.). The entity recognition strategy engine analyzes the business feedback information, identifies key entities with business meaning, such as order number, product name, amount, date, etc., and combines the identified entities and their types into entity information.

[0021] Furthermore, a domain knowledge graph is pre-constructed, containing a large number of entity nodes (such as "phone", "order", "refund") and their relationships and attribute information. Using entity information (such as the entity "phone") as a query condition, the entity node is located in the domain knowledge graph, and its associated attribute information is extracted, including the inherent characteristics of the entity information (such as the brand and model of "phone"), common problems (such as "phone lag" and "battery not lasting"), and user behavior patterns (such as "paying attention to performance after six months of use").

[0022] Furthermore, a list of "Frequently Asked Questions" (e.g., "How to solve phone lag" and "Do I need to replace my phone?") is extracted from the attribute information. Each frequently asked question corresponds to a potential implicit intent (e.g., "Seeking performance optimization solutions" and "Understanding trade-in policies"). The implicit intents are used to form a set of candidate intents. The semantic similarity between the business feedback information and each candidate intent is calculated, and the candidate intent with the highest similarity is taken as the implicit intent.

[0023] Furthermore, explicit and implicit intentions are structurally combined to form user feedback intentions that contain multiple layers of semantics.

[0024] S2. Obtain historical business behavior data including the target user's consumption amount, consumption frequency, and complaint records, and extract the consumption characteristics, credit characteristics, and activity characteristics of the historical business behavior data, and construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics.

[0025] In this embodiment of the invention, the historical business behavior data refers to all recordable behavioral trajectories and business operation records generated by the target user during the interaction with the platform over a period of time. The user profile includes multi-dimensional features such as user value stratification, credit rating, behavioral preferences, and group affiliation.

[0026] Among them, historical business behavior data can include consumption data such as consumption amount and consumption frequency, interaction data such as login frequency and login time distribution, service record data such as historical complaint records and consultation records, and risk behavior data such as overdue records and dispute records.

[0027] In this embodiment of the invention, during the process of extracting the consumption characteristics, credit characteristics, and activity characteristics of the historical business behavior data, the cumulative consumption amount, average consumption amount, and consumption frequency of the target user are calculated based on the historical business behavior data. The cumulative consumption amount, average consumption amount, and consumption frequency are combined into consumption characteristics. The average daily login frequency and average dwell time of the target user are statistically analyzed based on the historical business behavior data. The average daily login frequency and average dwell time are combined into activity characteristics.

[0028] Furthermore, based on historical business behavior data, the number of historical complaints and refund requests of target users are statistically analyzed, and the number of historical complaints and refund requests are combined into a risk dimension. Based on historical business behavior data, the number of reviews and positive review rate of target users are statistically analyzed, and the number of reviews and positive review rate are combined into a social dimension. The risk dimension and the social dimension are combined into a credit feature.

[0029] In this embodiment of the invention, constructing a user profile based on the consumption characteristics, the credit characteristics, and the activity characteristics includes: Based on the consumption characteristics, a value level analysis is performed on the target user to obtain the value level of the target user; Based on the credit characteristics, the credit rating of the target user is analyzed to obtain the credit rating of the target user; Based on the activity characteristics, the activity level of the target user is analyzed to obtain the activity level of the target user; A user profile is generated based on the value level, the credit level, and the activity level.

[0030] In this embodiment of the invention, duplicate records can be removed from historical business behavior data, missing values ​​can be processed (such as filling with default values ​​or deleting invalid records), and outliers can be identified and corrected. The historical business behavior data after the cleaning operation is unified in data format, and finally the numerical data in the unified format historical business behavior data is normalized to form standard behavior data.

[0031] Furthermore, in the process of analyzing the value level of the target user based on the consumption characteristics to obtain the value level of the target user, the consumption characteristics are transformed into consumption feature vectors, and the consumption feature vectors are input into a pre-trained value level classification model (such as an RFM scoring model based on K-Means clustering or a multi-class logistic regression model), and the model outputs high / medium / low value levels; then the credit characteristics are transformed into credit feature vectors, and input into a credit scoring card model (such as a scoring card based on logistic regression), the credit scoring card model calculates the total credit score, and then maps the total credit score to a preset score range (including excellent / good / medium / poor / very poor credit levels) to obtain the credit level; the credit characteristics are transformed into activity feature vectors, and the activity feature vectors are input into an activity level classifier based on threshold rules or clustering algorithms (such as K-Means dividing users into high, medium, and low clusters), and the activity level is output; the three value levels, the credit level, and the activity level are combined into a structured data object as the user profile of the target user.

[0032] In this embodiment of the invention, the value level classification model automatically groups users into three clusters, each cluster corresponding to a value level (high value, medium value, low value), and calculates the center point of each cluster (i.e., the average value of the feature vectors of all users in that cluster, representing the typical characteristics of users at that level). For example, user Zhang San's total spending in the past 30 days is 5,000 yuan, with 5 purchases, an average order value of 1,000 yuan, and a recent purchase interval of 2 days. These values ​​are combined to form a consumption feature vector [5000, 5, 1000, 2]. The consumption feature vector is input into a pre-trained value level model. The model calculates the distance between the consumption feature vector and the center point of each cluster, determines that it is closest to the center of the high-value user cluster, and outputs high value as Zhang San's value level.

[0033] S3. Identify the user type tags of the target user, and adjust the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights.

[0034] In this embodiment of the invention, the user type tag refers to an identifier used to describe the category or group attribute to which a user belongs, including sensitive users, non-sensitive users, high-value users, and growth-oriented users.

[0035] Sensitive users refer to users with high-risk behavioral characteristics, high-value attributes, or potential churn tendency, while non-sensitive users refer to users with low credit ratings, normal behavioral characteristics, and no need for complex intent analysis.

[0036] In this embodiment of the invention, the user type tag for identifying the target user includes: The historical business behavior data is segmented according to a preset time window to obtain multiple time window sub-sequences; Extract sub-activity features, sub-consumption features, and sub-complaint features from each time window subsequence, and aggregate the sub-activity features, sub-consumption features, and sub-complaint features into behavioral pattern features; The behavioral pattern features are transformed into behavioral pattern vectors; Calculate the similarity between the behavior pattern vector and each user type in a preset user type cluster containing user group categories, and select the user category with the highest similarity from the similarity as a candidate user type label; The candidate user type labels are subjected to time-series consistency verification to obtain user type labels.

[0037] In this embodiment of the invention, starting from the earliest historical business behavior data, the historical business behavior data is divided into multiple non-overlapping or partially overlapping time segments according to a preset time window (such as seven days or thirty days as a time window), and each time segment is a time window subsequence.

[0038] Furthermore, features such as total number of logins, total consumption amount, consumption frequency, number of complaints, and number of active days are extracted within each time window subsequence. The total consumption amount and consumption frequency are combined into sub-consumption features, the total number of logins and number of active days are combined into sub-activity features, and the number of complaints is used as a sub-complaint feature. The sub-activity features, the sub-consumption features, and the sub-complaint features are arranged in a preset order to obtain behavioral pattern features. The non-numerical features in the behavioral pattern features are encoded to generate a behavioral pattern vector.

[0039] In this embodiment of the invention, the preset user type cluster containing user group categories refers to multiple group categories divided after unsupervised learning of the behavioral pattern vectors of a large number of historical users. Each category represents a user group with similar behavioral characteristics and is assigned an interpretable label (such as high-value users, potential churned users, etc.).

[0040] Furthermore, the similarity between the behavior pattern vector and each user type in the preset user type cluster is calculated, and the user type corresponding to the highest similarity is selected as the candidate user type label corresponding to the subsequence of the time window.

[0041] Furthermore, the candidate user type labels corresponding to each time window subsequence are traversed, and it is compared whether the candidate user type labels corresponding to the current time window subsequence are the same as those corresponding to the previous and next time window subsequences. If the difference between the user type label of a certain time window subsequence and the user type labels of the preceding and following time window subsequences exceeds the preset acceptable range, the user type label of the current time window subsequence is regarded as an isolated abnormal label and is removed. After removal, the frequency of each user type label in the remaining user type labels is counted, and the user type label with the highest frequency is selected as the user type label of the target user.

[0042] In this embodiment of the invention, the step of scaling and adjusting the preset initial tag weights of the user type tags based on the activity characteristics in the user profile to obtain the tag weights includes: The activity level corresponding to the activity feature in the user profile is mapped to a preset activity level mapping table to obtain the activity level; The first adjustment weight is obtained by performing a linear operation between the preset initial tag weight corresponding to the user type tag and the activity level; Query the credit penalty coefficient corresponding to the credit rating in the user profile from the preset credit penalty mapping table; The second adjustment weight is obtained by performing a difference analysis between the first adjustment weight and the credit penalty coefficient. The second adjustment weight is normalized to obtain the label weight.

[0043] In this embodiment of the invention, the activity level includes high activity, medium activity, and low activity. The activity level is mapped to a numerical value according to a preset activity level mapping table, and the mapped numerical value is used as the activity level. For example, the high activity level corresponds to a larger numerical value (such as 1.2), the low activity level corresponds to a smaller numerical value (such as 0.8), and the medium activity level corresponds to a medium numerical value (such as 1.0).

[0044] Furthermore, based on historical experience, preset initial tag weights are set for user type tags, and the preset initial tag weights are multiplied by the activity level to obtain the first adjustment weight.

[0045] In this embodiment of the invention, the preset credit penalty mapping table is configured with a corresponding penalty coefficient for each credit level (such as excellent, good, medium, poor, and very poor). The worse the credit level, the larger the penalty coefficient.

[0046] Furthermore, the first adjustment weight is subtracted from the credit penalty coefficient to obtain the second adjustment weight; the third adjustment weight is normalized using the min-max normalization method, and the normalized third adjustment weight is used as the label weight.

[0047] In this embodiment of the invention, the historical business behavior data is segmented according to a preset time window, and the user's long-term behavior is decomposed into a phased sequence. This facilitates the capture of the dynamic trend of user behavior changes over time and avoids the one-sidedness of a single statistical analysis. By mapping the behavior pattern vector to a preset user type cluster, clustering matching is used to quickly locate the user's potential type tendency at each stage, providing a preliminary judgment basis for subsequent verification. By performing temporal consistency verification on the candidate user type labels, misjudged labels caused by a single abnormal behavior are filtered out, ensuring that the final labels can truly and stably reflect the user's long-term type characteristics.

[0048] S4. Adjust the label weights by attenuation based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights, and use the attenuated label weights as the target user type feature values ​​of the target user.

[0049] In this embodiment of the invention, the step of adjusting the label weights based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights includes: The consumption and credit characteristics in the user profile are transformed into a first feature vector; Each feature in the first feature vector is weighted by the label weight to obtain the second feature vector; Extract the most recent active timestamp of the target user from the user profile, obtain the current system timestamp, and calculate the time interval between the active timestamp and the system timestamp; Substituting the time interval into a preset decay function formula, the time decay factor of the target user is obtained; The second feature vector is linearly operated on with the time decay factor to obtain the decayed label weight.

[0050] In this embodiment of the invention, these values ​​are arranged one by one according to a predefined fixed dimensional order (e.g., consumption features first, then credit features) to form a one-dimensional list of values, namely the first feature vector.

[0051] Furthermore, each feature vector in the first feature vector is multiplied by the label weight to obtain the second feature vector.

[0052] In this embodiment of the invention, the time decay factor is used to quantify the degree of value decay of user historical behavior data over time. That is, the closer the behavior is to the current time, the higher its reference value for the current decision, and the more distant the behavior is, the lower its value.

[0053] Furthermore, the last active timestamp of the target user (such as the time of the last login, the last transaction, or the last submission of feedback) is extracted from the user profile. The time interval between the last active timestamp and the current system time is calculated. This time interval is then substituted into a preset decay function formula to calculate the time decay factor of the target user.

[0054] The preset attenuation function formula can be as follows: Right now This can represent the time decay factor, with a value range of (0, 1]. This represents the natural constant, approximately 2.718. Indicates time interval, This represents the attenuation rate, with a value range of [0.01 ~ 0.1].

[0055] Furthermore, each feature vector in the second feature vector is multiplied by the time decay factor, and the results of the multiplication are added together to obtain the decayed label weight, which is the target user type feature value.

[0056] In this embodiment of the invention, each feature vector in the first feature vector is weighted with the label weight, which can reflect the difference in the degree of influence of different user type labels on the target user type, and make the calculation results more in line with the business's emphasis on different labels.

[0057] S5. Determine whether the target user type feature value exceeds a preset feature threshold.

[0058] In this embodiment of the invention, the target user type feature value is compared with a preset feature threshold to determine whether the target user type feature value exceeds the preset feature threshold.

[0059] S6. When the target user type feature value does not exceed the feature threshold, the business feedback information is assigned to manual processing.

[0060] In this embodiment of the invention, business feedback information and related context (such as user ID, historical business behavior data summary) are encapsulated into a work order to be processed, and the work order to be processed is inserted into the human customer service processing queue according to a preset priority strategy (such as according to the order of submission time, the urgency of the problem type) to wait for allocation. The human customer service system takes the work order to be processed from the human customer service processing queue and assigns it to online customer service personnel for manual review and processing.

[0061] S7. When the target user type feature value exceeds the preset feature threshold, the matching degree of the user feedback intent is calculated with each business processing strategy in the preset business strategy library used to process business feedback information. The business processing strategy with a matching degree greater than the preset matching threshold is selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategy.

[0062] In this embodiment of the invention, the step of calculating the matching degree between the user feedback intent and each business processing strategy in a preset business strategy library used for processing business feedback information, and selecting business processing strategies from the business strategy library whose matching degree is greater than a preset matching threshold, includes: The matching degree of the user feedback intent is calculated by matching it with each business processing strategy in the preset business strategy library to obtain the matching degree corresponding to each business processing strategy. The business processing strategies with a matching degree greater than a preset matching threshold are aggregated into a candidate business processing strategy set; The value level, credit level, and activity level of each business processing strategy in the candidate business processing strategy set are analyzed. The similarity between the analyzed value level, credit level, and activity level and the value level, credit level, and activity level in the user profile is calculated to obtain the strategy similarity corresponding to each candidate business processing strategy set. Based on the strategy similarity, the candidate business processing strategy set is sorted in descending order into a sorted business strategy set, and the top N business strategies are selected from the sorted business strategy set. Extract the business operation instructions corresponding to the selected business strategies, and detect whether there are logical conflicts between the business operation instructions; If there is a logical conflict between the business operation instructions, the business strategy with the highest priority is selected according to the preset priority of the business strategy, and the selected business strategy is used as the business processing strategy.

[0063] In this embodiment of the invention, the preset business strategy library contains multiple business processing strategies. The business processing strategy refers to a set of predefined executable operation rules for specific user intentions or business scenarios, such as specific processing solutions such as automatic refund, issuing coupons, and transferring to manual review. Each business processing strategy has preset level conditions that users applying the strategy must meet (for example, strategy A requires the user's value level to be "high value", credit level to be "good", and activity level to be "high activity").

[0064] Furthermore, the user value level, credit level, activity level, and other conditions of each business processing strategy are analyzed. The value level, credit level, and activity level in the user profile are compared with the value level, credit level, and activity level required by each business processing strategy in a dimension-by-dimensional manner. That is, if a certain level dimension is completely consistent, the score is 1. If they are inconsistent, a partial matching score between 0 and 1 is given according to the level difference (such as the difference between "high value" and "medium value" being one level). The scores of each level dimension are weighted and summed to obtain the strategy similarity corresponding to each candidate business strategy.

[0065] Furthermore, the selected business strategies are analyzed one by one. Based on the analysis results, the business operation instructions corresponding to each business strategy are extracted. The logical conflicts between the business operation instructions are checked. For business strategies with conflicts, the business strategy with higher priority among the conflicting business strategies is selected according to the preset priority of the business strategies. The business strategies without conflicts are merged with the selected business strategies to form a business processing strategy, which is then used for subsequent actual execution.

[0066] It can also perform resource availability verification on business processing strategies, perform condition verification on intermediate business processing strategies that have passed resource availability verification, and perform time verification on business processing strategies that have passed condition verification to obtain verified business processing strategies.

[0067] Taking a user's refund request as an example, the generated business processing strategy is "refund 20 yuan and issue a 10 yuan coupon to order A". In the resource availability verification stage, it is identified that the 10 yuan coupon is in short supply, so the coupon is automatically adjusted to an 8 yuan coupon that is still in stock, generating the intermediate business processing strategy "refund 20 yuan and issue an 8 yuan coupon". Then, condition verification is performed, and it is identified that the status of order A is normal and the user account is not frozen, so the condition verification is passed. Then, time verification is performed to confirm that the 8 yuan coupon is still valid, so the time verification is passed. After three steps of verification, the executable business processing strategy "refund 20 yuan and issue an 8 yuan coupon" is output.

[0068] As can be seen, the above solution proposes a business processing method and system based on user profiles, which aims to solve the pain points of misjudging the true intent of users, failing to provide differentiated services to high-value and potentially churned users, and reducing the accuracy of business processing. The core of the solution lies in receiving business feedback information submitted by target users to the business platform, performing semantic analysis on the business feedback information, classifying the business feedback information according to the analysis results to obtain the user feedback intent; acquiring historical business behavior data including the target user's consumption amount, consumption frequency, and complaint records, and extracting consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data, constructing a user profile based on the consumption characteristics, credit characteristics, and activity characteristics; identifying the target user's user type tags, scaling and adjusting the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain tag weights; attenuating the tag weights according to the consumption characteristics and credit characteristics in the user profile to obtain attenuated tag weights, and using the attenuated tag weights as the target user type feature value of the target user; when the target user type feature value exceeds a preset feature threshold, calculating the matching degree between the user feedback intent and each business processing strategy in a preset business strategy library used to process business feedback information, filtering out business processing strategies with a matching degree greater than the preset matching threshold from the business strategy library, and processing the business feedback information according to the filtered business processing strategies.

[0069] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0070] In one embodiment, a user profile-based service processing apparatus is provided, which corresponds one-to-one with the user profile-based service processing methods described in the above embodiments. For example... Figure 3 As shown, the user profile-based business processing device includes a user feedback intent classification module 101, a user profile construction module 102, a user type label recognition module 103, a label weight adjustment module 104, a target user type feature value adjustment module 105, and a business feedback information processing module 106. Detailed descriptions of each functional module are as follows: The user feedback intent classification module 101 is used to receive business feedback information submitted by the target user to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information according to the analysis results, and obtain the user feedback intent. The user profile building module 102 is used to acquire historical business behavior data including the target user's consumption amount, consumption frequency, and complaint records, and extract the consumption characteristics, credit characteristics, and activity characteristics of the historical business behavior data, and build a user profile based on the consumption characteristics, the credit characteristics, and the activity characteristics. User type tag recognition module 103 is used to recognize the user type tag of the target user; The tag weight adjustment module 104 is used to scale and adjust the preset initial tag weight of the user type tag according to the activity characteristics in the user profile to obtain the tag weight; The target user type feature value adjustment module 105 is used to adjust the label weights by attenuation according to the consumption characteristics and credit characteristics in the user profile, to obtain the attenuated label weights, and to use the attenuated label weights as the target user type feature values ​​of the target user. The business feedback information processing module 106 is used to calculate the matching degree between the user feedback intent and each business processing strategy in the preset business strategy library for processing business feedback information when the target user type feature value exceeds the preset feature threshold, filter out the business processing strategies with matching degree greater than the preset matching threshold from the business strategy library, and process the business feedback information according to the filtered business processing strategies.

[0071] In one embodiment, the user feedback intent classification module 101, when performing the intent classification of the business feedback information based on the analysis results to obtain the user feedback intent, is specifically used for: Based on the analysis results, the business feedback information is mapped to a preset set of intent categories with probability values; The intent category with the highest probability value in the intent category set is selected as the explicit intent, and the highest probability value is used as the confidence level of the explicit intent; Determine whether the confidence level is lower than a preset confidence threshold; When the confidence level is not lower than the confidence threshold, the explicit intent is taken as the user feedback intent. When the confidence level is lower than the confidence level threshold, the business feedback information is matched with the preset entity recognition strategy engine to obtain entity information; The attribute information of the entity is queried from a preset domain knowledge graph, and the attribute information includes common business issues; A candidate intent set is generated based on common issues in the attribute information. The similarity between the business feedback information and each candidate intent in the candidate intent set is calculated. The candidate intent with the highest similarity is selected as the implicit intent. The explicit intent and the implicit intent are combined into a user feedback intent.

[0072] In one embodiment, the user profile building module 102, when building a user profile based on the consumption characteristics, the credit characteristics, and the activity characteristics, is specifically used for: Based on the consumption characteristics, a value level analysis is performed on the target user to obtain the value level of the target user; Based on the credit characteristics, the credit rating of the target user is analyzed to obtain the credit rating of the target user; Based on the activity characteristics, the activity level of the target user is analyzed to obtain the activity level of the target user; A user profile is generated based on the value level, the credit level, and the activity level.

[0073] In one embodiment, the user type tag recognition module 103, when recognizing the user type tag of the target user, is specifically used for: The historical business behavior data is segmented according to a preset time window to obtain multiple time window sub-sequences; Extract sub-activity features, sub-consumption features, and sub-complaint features from each time window subsequence, and arrange the sub-activity features, sub-consumption features, and sub-complaint features in a preset order to form behavioral pattern features; The behavioral pattern features are transformed into behavioral pattern vectors; Calculate the similarity between the behavior pattern vector and each user type in a preset user type cluster containing user group categories, and select the user category with the highest similarity from the similarity as a candidate user type label; The candidate user type labels are subjected to time-series consistency verification to obtain user type labels.

[0074] In one embodiment, the tag weight adjustment module 104, when scaling and adjusting the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights, is specifically used for: The activity level corresponding to the activity feature in the user profile is mapped to a preset activity level mapping table to obtain the activity level; The first adjustment weight is obtained by performing a linear operation between the preset initial tag weight corresponding to the user type tag and the activity level; Query the credit penalty coefficient corresponding to the credit rating in the user profile from the preset credit penalty mapping table; The second adjustment weight is obtained by performing a difference analysis between the first adjustment weight and the credit penalty coefficient. The second adjustment weight is normalized to obtain the label weight.

[0075] In one embodiment, the target user type feature value adjustment module 105, when performing the attenuation adjustment of the tag weights based on the consumption characteristics and credit characteristics in the user profile to obtain the attenuated tag weights, is specifically used for: The consumption and credit characteristics in the user profile are transformed into a first feature vector; Each feature in the first feature vector is weighted by the label weight to obtain the second feature vector; Extract the most recent active timestamp of the target user from the user profile, obtain the current system timestamp, and calculate the time interval between the active timestamp and the system timestamp; Substituting the time interval into a preset decay function formula, the time decay factor of the target user is obtained; The second feature vector is linearly operated on with the time decay factor to obtain the decayed label weight.

[0076] In one embodiment, the business feedback information processing module 106, when performing the step of calculating the matching degree between the user feedback intent and each business processing strategy in a preset business strategy library used for processing business feedback information, and filtering out business processing strategies from the business strategy library whose matching degree is greater than a preset matching threshold, specifically performs the following: The matching degree of the user feedback intent is calculated by matching it with each business processing strategy in the preset business strategy library to obtain the matching degree corresponding to each business processing strategy. The business processing strategies with a matching degree greater than a preset matching threshold are aggregated into a candidate business processing strategy set; The value level, credit level, and activity level of each business processing strategy in the candidate business processing strategy set are analyzed. The similarity between the analyzed value level, credit level, and activity level and the value level, credit level, and activity level in the user profile is calculated to obtain the strategy similarity corresponding to each candidate business processing strategy set. Based on the strategy similarity, the candidate business processing strategy set is sorted in descending order into a sorted business strategy set, and the top N business strategies are selected from the sorted business strategy set. Extract the business operation instructions corresponding to the selected business strategies, and detect whether there are logical conflicts between the business operation instructions; If there is a logical conflict between the business operation instructions, the business strategy with the highest priority is selected according to the preset priority of the business strategy, and the selected business strategy is used as the business processing strategy.

[0077] This invention provides a business processing device based on user profiles, which aims to solve the pain points of misjudging the true intent of users, failing to provide differentiated services to high-value and potentially churned users, and reducing the accuracy of business processing. The core of the solution lies in receiving business feedback information submitted by target users to the business platform, performing semantic analysis on the business feedback information, classifying the business feedback information according to the analysis results to obtain the user feedback intent; acquiring historical business behavior data including the target user's consumption amount, consumption frequency, and complaint records, and extracting consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data, constructing a user profile based on the consumption characteristics, credit characteristics, and activity characteristics; identifying the target user's user type tags, scaling and adjusting the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain tag weights; attenuating the tag weights according to the consumption characteristics and credit characteristics in the user profile to obtain attenuated tag weights, and using the attenuated tag weights as the target user type feature value of the target user; when the target user type feature value exceeds a preset feature threshold, calculating the matching degree between the user feedback intent and each business processing strategy in a preset business strategy library used to process business feedback information, filtering out business processing strategies with a matching degree greater than the preset matching threshold from the business strategy library, and processing the business feedback information according to the filtered business processing strategies.

[0078] Specific limitations regarding the user profile-based business processing device can be found in the limitations of the user profile-based business processing method described above, and will not be repeated here. Each module in the aforementioned user profile-based business processing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0079] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a user profile-based business processing method on the server side.

[0080] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements client-side functions or steps of a user profile-based business processing method.

[0081] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information according to the analysis results, and obtain the user feedback intent; Obtain historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data; construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. Identify the user type tags of the target user, and adjust the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights; The label weights are adjusted by attenuation based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights, and the attenuated label weights are used as the target user type feature values ​​of the target user. When the target user type feature value exceeds a preset feature threshold, the matching degree of the user feedback intent is calculated with each business processing strategy in the preset business strategy library used to process business feedback information. The business processing strategy with a matching degree greater than the preset matching threshold is selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategy.

[0082] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information according to the analysis results, and obtain the user feedback intent; Obtain historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data; construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. Identify the user type tags of the target user, and adjust the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights; The label weights are adjusted by attenuation based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights, and the attenuated label weights are used as the target user type feature values ​​of the target user. When the target user type feature value exceeds a preset feature threshold, the matching degree of the user feedback intent is calculated with each business processing strategy in the preset business strategy library used to process business feedback information. The business processing strategy with a matching degree greater than the preset matching threshold is selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategy.

[0083] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0084] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0085] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0086] Finally, it should be noted that if any software tools or components not belonging to this company appear in the embodiments of the application, they are merely illustrative examples and do not represent actual use. The embodiments described above are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A business processing method based on user profiles, characterized in that, include: Receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information according to the analysis results, and obtain the user feedback intent; Obtain historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data; construct a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. Identify the user type tags of the target user, and adjust the preset initial tag weights of the user type tags according to the activity characteristics in the user profile to obtain the tag weights; The label weights are adjusted by attenuation based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights, and the attenuated label weights are used as the target user type feature values ​​of the target user. When the target user type feature value exceeds a preset feature threshold, the matching degree of the user feedback intent is calculated with each business processing strategy in the preset business strategy library used to process business feedback information. The business processing strategy with a matching degree greater than the preset matching threshold is selected from the business strategy library, and the business feedback information is processed according to the selected business processing strategy.

2. The user profile-based business processing method as described in claim 1, characterized in that, The step of constructing a user profile based on the consumption characteristics, credit characteristics, and activity characteristics includes: Based on the consumption characteristics, a value level analysis is performed on the target user to obtain the value level of the target user; Based on the credit characteristics, the credit rating of the target user is analyzed to obtain the credit rating of the target user; Based on the activity characteristics, the activity level of the target user is analyzed to obtain the activity level of the target user; A user profile is generated based on the value level, the credit level, and the activity level.

3. The user profile-based business processing method as described in claim 1, characterized in that, The step of scaling and adjusting the preset initial tag weights of the user type tags based on the activity characteristics in the user profile to obtain the tag weights includes: The activity level corresponding to the activity feature in the user profile is mapped to a preset activity level mapping table to obtain the activity level; The first adjustment weight is obtained by performing a linear operation between the preset initial tag weight corresponding to the user type tag and the activity level; Query the credit penalty coefficient corresponding to the credit rating in the user profile from the preset credit penalty mapping table; The second adjustment weight is obtained by performing a difference analysis between the first adjustment weight and the credit penalty coefficient. The second adjustment weight is normalized to obtain the label weight.

4. The user profile-based business processing method as described in claim 1, characterized in that, The step of adjusting the label weights based on the consumption and credit characteristics in the user profile to obtain the attenuated label weights includes: The consumption and credit characteristics in the user profile are transformed into a first feature vector; Each feature in the first feature vector is weighted by the label weight to obtain the second feature vector; Extract the most recent active timestamp of the target user from the user profile, obtain the current system timestamp, and calculate the time interval between the active timestamp and the system timestamp; Substituting the time interval into a preset decay function formula, the time decay factor of the target user is obtained; The second feature vector is linearly operated on with the time decay factor to obtain the decayed label weight.

5. The user profile-based business processing method as described in claim 1, characterized in that, The user type label for identifying the target user includes: The historical business behavior data is segmented according to a preset time window to obtain multiple time window sub-sequences; Extract sub-activity features, sub-consumption features, and sub-complaint features from each time window subsequence, and arrange the sub-activity features, sub-consumption features, and sub-complaint features in a preset order to form behavioral pattern features; The behavioral pattern features are transformed into behavioral pattern vectors; Calculate the similarity between the behavior pattern vector and each user type in a preset user type cluster containing user group categories, and select the user category with the highest similarity from the similarity as a candidate user type label; The candidate user type labels are subjected to time-series consistency verification to obtain user type labels.

6. The business processing method based on user profiles as described in claim 1, characterized in that, The step of classifying the business feedback information according to the analysis results to obtain the user feedback intent includes: Based on the analysis results, the business feedback information is mapped to a preset set of intent categories with probability values; The intent category with the highest probability value in the intent category set is selected as the explicit intent, and the highest probability value is used as the confidence level of the explicit intent; Determine whether the confidence level is lower than a preset confidence threshold; When the confidence level is not lower than the confidence threshold, the explicit intent is taken as the user feedback intent. When the confidence level is lower than the confidence level threshold, the business feedback information is matched with the preset entity recognition strategy engine to obtain entity information; The attribute information of the entity is queried from a preset domain knowledge graph, and the attribute information includes common business issues; A candidate intent set is generated based on common issues in the attribute information. The similarity between the business feedback information and each candidate intent in the candidate intent set is calculated. The candidate intent with the highest similarity is selected as the implicit intent. The explicit intent and the implicit intent are combined into a user feedback intent.

7. The business processing method based on user profiles as described in claim 1, characterized in that, The step of calculating the matching degree between the user feedback intent and each business processing strategy in the preset business strategy library used for processing business feedback information, and filtering out business processing strategies with a matching degree greater than a preset matching threshold from the business strategy library, includes: The matching degree of the user feedback intent is calculated by matching it with each business processing strategy in the preset business strategy library to obtain the matching degree corresponding to each business processing strategy. The business processing strategies with a matching degree greater than a preset matching threshold are aggregated into a candidate business processing strategy set; The value level, credit level, and activity level of each business processing strategy in the candidate business processing strategy set are analyzed. The similarity between the analyzed value level, credit level, and activity level and the value level, credit level, and activity level in the user profile is calculated to obtain the strategy similarity corresponding to each candidate business processing strategy set. Based on the strategy similarity, the candidate business processing strategy set is sorted in descending order into a sorted business strategy set, and the top N business strategies are selected from the sorted business strategy set. Extract the business operation instructions corresponding to the selected business strategies, and detect whether there are logical conflicts between the business operation instructions; If there is a logical conflict between the business operation instructions, the business strategy with the highest priority is selected according to the preset priority of the business strategy, and the selected business strategy is used as the business processing strategy.

8. A business processing device based on user profiles, characterized in that, include: The user feedback intent classification module is used to receive business feedback information submitted by target users to the business platform, perform semantic analysis on the business feedback information, classify the business feedback information into intents based on the analysis results, and obtain the user feedback intent. The user profile building module is used to acquire historical business behavior data containing the target user's consumption amount, consumption frequency, and complaint records, and extract consumption characteristics, credit characteristics, and activity characteristics from the historical business behavior data, and build a user profile based on the consumption characteristics, credit characteristics, and activity characteristics. The user type tag recognition module is used to identify the user type tag of the target user; The tag weight adjustment module is used to scale and adjust the preset initial tag weight of the user type tag according to the activity characteristics in the user profile to obtain the tag weight; The target user type feature value adjustment module is used to adjust the label weights by attenuation based on the consumption characteristics and credit characteristics in the user profile, to obtain the attenuated label weights, and to use the attenuated label weights as the target user type feature values ​​of the target user. The business feedback information processing module is used to calculate the matching degree between the user feedback intent and each business processing strategy in the preset business strategy library for processing business feedback information when the target user type feature value exceeds the preset feature threshold, filter out the business processing strategies with matching degree greater than the preset matching threshold from the business strategy library, and process the business feedback information according to the filtered business processing strategies.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the user profile-based business processing method as described in any one of claims 1 to 7.

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

Images