User perception mapping method and system based on service experience model

Abstract

The application provides a user perception mapping method and system based on a service experience model, constructs a perception anchor point set of service experience features and user perception expressions, wherein the service experience features include service interaction fluency and content presentation fitness, and the user perception expressions are direct feeling descriptions of users on service use. Real-time service interaction data is analyzed for features, and real-time service experience features consistent with the service experience feature types in the perception anchor point set are extracted. The real-time service experience features are matched with the perception anchor point set by calling a service experience model, and a matching result is generated by dynamically adjusting the matching weight through an association reinforcement module. The matching result is used to perform perception expression conversion, and real-time user perception information is formed in combination with a real-time service scene. The real-time user perception information is compared with a preset perception optimization benchmark, and perception optimization guidance including a perception improvement direction and specific service adjustment suggestions is generated to provide accurate guidance for service optimization.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence, and more specifically, to a user perception mapping method and system based on a business experience model. Background Technology

[0002] In today's era of booming digital business, user experience is increasingly valued. Accurately understanding user perceptions of a service is crucial for improving service quality and enhancing user satisfaction and loyalty. However, existing methods for acquiring user perceptions have several limitations. On one hand, traditional user surveys often rely on proactive user feedback, which is not only time-consuming but also struggles to guarantee the comprehensiveness and accuracy of the feedback, as many users may be unable to express their true feelings promptly and accurately for various reasons. On the other hand, while some data-driven user perception assessment methods can collect operational data, they lack a mechanism to effectively link service experience characteristics with user subjective perceptions. This makes it difficult to deeply understand users' specific feelings and needs regarding the service, and thus fails to provide precise and targeted guidance for service optimization. Therefore, there is an urgent need for a method that can accurately and efficiently map service experience characteristics into user perception information to address the problems existing in current technologies. Summary of the Invention

[0003] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a user perception mapping method based on a business experience model, the method comprising: Construct a set of perception anchors for business experience features and user perception expressions. The business experience features include the smoothness of business interaction and the fit of content presentation. The user perception expressions include the direct feelings and descriptions of users when using the business. The real-time business interaction data is processed by feature parsing to extract real-time business experience features that are consistent with the business experience feature types in the set of perception anchors. The real-time business interaction data is dynamic data generated by users during the use of the business. The business experience model is invoked to match the real-time business experience features with the set of perception anchors. The matching weights are dynamically adjusted through the association enhancement module built into the model to generate anchor matching results. Based on the anchor point matching results, a perception representation transformation is performed. The user perception representation in the matching anchor point unit corresponding to the real-time business experience features is used as the basis, and detailed descriptions are added in combination with the real-time business scenario to form real-time user perception information. The real-time user perception information is compared with a preset perception optimization benchmark to generate perception optimization guidelines that include perception improvement directions and specific business adjustment suggestions.

[0004] In another aspect, embodiments of the present invention also provide a user perception mapping system based on a business experience model, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0005] Based on the above, this embodiment of the invention constructs a set of perception anchor points to accurately associate business experience features with user perception expressions. Feature parsing processing of real-time business interaction data accurately extracts real-time business experience features consistent with the business experience feature types in the perception anchor point set, ensuring data accuracy and consistency. Anchor point matching is performed using a business experience model, and the matching weights are dynamically adjusted through an association enhancement module, making the matching results more closely aligned with actual business scenarios, thus improving the accuracy and reliability of the matching. Perception expression transformation is performed based on the anchor point matching results, supplementing detailed descriptions with real-time business scenarios, resulting in more comprehensive and accurate real-time user perception information that truly reflects users' feelings about the business. Finally, the real-time user perception information is compared with a preset perception optimization benchmark. The generated perception optimization guidelines provide clear directions and specific suggestions for business optimization, helping to improve business quality, enhance user satisfaction, and thus gain an advantage in fierce market competition. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the execution flow of the user perception mapping method based on the business experience model provided in this embodiment of the invention.

[0007] Figure 2 This is a schematic diagram of exemplary hardware and software components of a user perception mapping system based on a business experience model provided in an embodiment of the present invention. Detailed Implementation

[0008] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a user perception mapping method based on a business experience model, as provided in an embodiment of the present invention. The following is a detailed description of this user perception mapping method based on a business experience model.

[0009] Step S110: Construct a set of perception anchor points for business experience features and user perception expressions. The business experience features include the smoothness of business interaction and the fit of content presentation. The user perception expressions include the user's direct feelings about using the business.

[0010] In this embodiment, this step is the foundation for all subsequent steps. It aims to establish a set of correspondences between business experience features and user perception descriptions, providing a reference for mapping real-time business experience features to user perception information.

[0011] Step S111: Collect historical interaction datasets and synchronized user feedback datasets during business usage.

[0012] In this embodiment, this step is used to obtain the raw data required to construct the perception anchor set. The historical interaction dataset is a collection of all interaction-related data generated by the user during past use of the service, such as user click operation records, page jump records, content browsing time records, etc.; the synchronous user feedback dataset is a collection of subjective user feedback data collected through the service's built-in feedback entry during or after the user's use of the service, such as text evaluations entered by the user in the feedback box, text descriptions corresponding to the selected satisfaction star rating, etc.

[0013] Step S112: Extract business interaction-related data and content presentation-related data from the historical interaction dataset, perform fluency feature quantification on the business interaction-related data to obtain business interaction fluency features, and perform fit feature extraction on the content presentation-related data to obtain content presentation fit features.

[0014] In this embodiment, this step is used to extract and quantify business experience features from the historical interaction dataset. Business interaction-related data refers to data in the historical interaction dataset directly related to the interaction between the user and the business, such as response time data for user clicks and page loading time data. Content presentation-related data refers to data in the historical interaction dataset related to the content presented to the user by the business, such as content title data, content category data, and the number of times the user clicked on the content. When quantifying the fluency features of the business interaction-related data, statistical analysis methods can be used, such as calculating the average, maximum, and minimum values ​​of the user click response time, and incorporating these statistics as components of the business interaction fluency features. When extracting the fit features of the content presentation-related data, association analysis methods can be used, such as analyzing the degree of matching between the content categories browsed by the user and the user's historical browsing preference categories, and incorporating this degree of matching as a component of the content presentation fit features.

[0015] Step S113: Perform semantic segmentation on the user subjective feelings text in the synchronous user feedback dataset, extract the core sentences that can directly reflect the user's user experience as the user perception expression, and remove text content that is ambiguous or irrelevant to the business experience.

[0016] In this embodiment, this step is used to extract effective user-perceived expressions from the synchronous user feedback dataset. Semantic segmentation can employ word segmentation techniques from natural language processing to break down the user's subjective feelings into individual words or phrases. Then, semantic understanding techniques are used to analyze the meaning of each word or phrase to determine whether it directly reflects the user's experience. For example, for the user input "This service is very smooth to use, and the content matches my preferences," after semantic segmentation, "very smooth to use" and "the content matches my preferences" can be extracted as the core statements; for the user input "The weather is really nice today," since it is irrelevant to the service experience, it will be discarded.

[0017] Step S114: Associate and bind the business interaction fluency characteristics and content presentation fit characteristics of the same period with the corresponding user perception expression to form an initial anchor unit. The initial anchor unit formed by association and binding each contains a set of business interaction fluency characteristics, content presentation fit characteristics and user perception expression.

[0018] In this embodiment, this step is used to associate business experience features with user-perceived expressions to form initial anchor units. The same time period refers to the business usage time period corresponding to the business interaction fluency feature and content presentation fit feature, which is the same as the business usage time period corresponding to the user-perceived expression. For example, if a user uses the service within a certain time period, generating corresponding business interaction fluency features and content presentation fit features, and submits a user-perceived expression within or after that time period, then this set of business interaction fluency features and content presentation fit features is associated and bound with the user-perceived expression to form an initial anchor unit.

[0019] Step S115: Perform duplicate removal on all initial anchor units, calculate the correlation stability between business experience features and user perception statements for each initial anchor unit, retain the initial anchor units whose correlation stability meets the preset standard, and classify and organize the retained initial anchor units according to business scenario type to form a set of perception anchors for business experience features and user perception statements.

[0020] In this embodiment, this step is used to screen and organize the initial anchor units to form the final set of perception anchors. Duplicate removal involves checking all initial anchor units; if two or more initial anchor units have identical business interaction fluency characteristics, content presentation fit characteristics, and user perception expressions, only one is retained. When calculating association stability, correlation analysis can be used to analyze the degree of correlation between business experience characteristics and user perception expressions. For example, the correlation coefficient between the response time statistic in the business interaction fluency characteristics and the "fluidity" related expressions in the user perception expressions is calculated, and this correlation coefficient is used as a measure of association stability. The preset standard is a pre-set association stability threshold, such as 0.6. If the association stability of an initial anchor unit is greater than or equal to this threshold, the initial anchor unit is retained; otherwise, it is removed. When classifying and organizing according to business scenario types, the business scenario type is a type based on the usage scenario of the business, such as shopping scenarios, information browsing scenarios, social scenarios, etc. Initial anchor units belonging to the same business scenario type are grouped into one category.

[0021] Step S1151: Perform feature hashing on all initial anchor units, generate a unique feature hash value for each initial anchor unit, identify duplicate initial anchor units by comparing feature hash values, retain one duplicate unit, delete the other duplicate units, and complete the duplicate item removal process.

[0022] In this embodiment, duplicate item removal is implemented. Feature hashing is a method that converts data into fixed-length hash values. A unique feature hash value is generated by performing hash calculations on the business interaction fluency features, content presentation fit features, and user perception expressions in the initial anchor unit. When comparing feature hash values, if two initial anchor units have the same feature hash value, they are considered duplicates; one is retained, and the other duplicate units are deleted.

[0023] Step S1152: Assign a unique identifier to each of the initial anchor units after duplicate item removal, extract the business experience feature data sequence and the corresponding user perception expression semantic vector from the processed initial anchor units, and input both into the association stability calculation module.

[0024] In this embodiment, the unique identifier is a unique identifier, such as a string or number, assigned to each initial anchor unit after duplicates have been removed, to distinguish different initial anchor units. The business experience feature data sequence is a sequence formed by arranging the business interaction fluency features and content presentation fit features in the initial anchor units in a certain order; the user-perceived expression semantic vector is a vector form converted from the user-perceived expression, for example, by converting the words in the user-perceived expression into vectors using word embedding technology, and then combining these vectors to obtain the user-perceived expression semantic vector. The association stability calculation module is a module used to calculate the association stability between business experience features and user-perceived expressions.

[0025] Step S1153: The association stability calculation module calculates the degree of association between the business experience feature data sequence and the user-perceived expression semantic vector through the semantic association analysis algorithm, and generates an association stability value. When the association stability value reaches the preset stability standard, the association between the business experience feature data sequence and the user-perceived expression semantic vector meets the requirements of actual application.

[0026] In this embodiment, a correlation stability value is calculated. The semantic correlation analysis algorithm is used to analyze the degree of correlation between two semantic objects. This algorithm can calculate the tightness of the correlation between the business experience feature data sequence and the user-perceived expression semantic vector, generating a correlation stability value. The preset stability standard is a pre-set threshold value for the correlation stability value, for example, 0.7. If the correlation stability value is greater than or equal to this threshold, the correlation between the business experience feature data sequence and the user-perceived expression semantic vector is considered to meet the requirements of practical application.

[0027] Step S1154: Set a preset standard for correlation stability, compare the correlation stability values ​​of the processed initial anchor units one by one with the preset standard, retain the initial anchor units whose correlation stability values ​​reach or exceed the preset standard, and remove the initial anchor units whose values ​​are lower than the preset standard.

[0028] In this embodiment, the initial anchor unit is selected. The association stability preset standard is a pre-set threshold value for the association stability of the initial anchor unit, for example, 0.6. The association stability value of each initial anchor unit is compared with the preset standard. If the association stability value reaches or exceeds the preset standard, the initial anchor unit is retained; otherwise, the initial anchor unit is discarded.

[0029] Step S1155: Extract the business scenario information corresponding to the retained initial anchor units, classify the retained initial anchor units according to the business scenario information, and group the initial anchor units belonging to the same business scenario into one category to form a scenario anchor group.

[0030] In this embodiment, the retained initial anchor units are classified. Business scenario information refers to the business usage scenario corresponding to each initial anchor unit, such as shopping scenario information, information browsing scenario information, etc. When classifying the retained initial anchor units according to business scenario information, initial anchor units with the same business scenario information are grouped together to form scenario anchor groups.

[0031] Step S1156: Perform feature consistency verification on each initial anchor unit in the divided scene anchor group to ensure that the business experience feature type of the initial anchor unit in the same group is consistent, and integrate all scene anchor groups to form a set of perception anchors that represent business experience features and user perception expressions.

[0032] In this embodiment, features are used to form the final set of perceptual anchor points. Feature consistency verification checks whether the business experience feature types of the initial anchor point units within the same scene anchor point group are consistent. For example, whether the types of business interaction fluency features are the same, or whether the types of content presentation fit features are the same. If the business experience feature types of the initial anchor point units within the same group are inconsistent, they are adjusted to make them consistent. Integrating all scene anchor point groups forms the set of perceptual anchor points that represent both business experience features and user perception expressions.

[0033] Step S120: Perform feature parsing processing on the real-time business interaction data, and extract real-time business experience features that are consistent with the business experience feature types in the perception anchor set. The real-time business interaction data is dynamic data generated during the user's use of the business.

[0034] In this embodiment, this step is used to extract real-time business experience features from real-time business interaction data that are consistent with the business experience feature types in the perception anchor set.

[0035] Step S121: Obtain real-time business interaction data generated during user use of services through the business data acquisition interface. The real-time business interaction data includes business response status data, content loading status data, and user operation trigger data.

[0036] In this embodiment, this step is used to acquire real-time business interaction data. The business data acquisition interface is an interface in the business system used to collect real-time business interaction data. Through this interface, various data generated during user interaction can be acquired in real time. Business response status data is data related to the business's response to user operations, such as the response time data after a user clicks an operation; content loading status data is loading-related data when the business presents content to the user, such as the page content loading time data; user operation trigger data is data related to user-triggered operations, such as operation record data of a user clicking a button.

[0037] Step S122: Perform data diversion processing on the real-time business interaction data, separate the business response status data and content loading status data, and transmit them to the corresponding feature extraction channels respectively. User operation trigger data is temporarily stored as auxiliary data.

[0038] In this embodiment, this step is used to perform data splitting processing on real-time business interaction data. Data splitting processing divides real-time business interaction data into different data streams according to the data type. For example, business response status data is divided into one data stream, content loading status data into another data stream, and user operation trigger data into a third data stream. The corresponding feature extraction channels are used to extract business interaction fluency features and content presentation fit features. Business response status data is transmitted to the channel used to extract business interaction fluency features, and content loading status data is transmitted to the channel used to extract content presentation fit features; user operation trigger data is temporarily stored as auxiliary data.

[0039] Step S123: In the business response feature extraction channel, perform time series feature extraction on the business response status data, capture the continuous change features of the business response, and convert the captured features into real-time business interaction smoothness features that are consistent with the business interaction smoothness feature format in the perception anchor set.

[0040] In this embodiment, this step is used to extract real-time business interaction fluency features. The business response feature extraction channel is used to extract business interaction fluency features. When extracting time-series features from business response status data, time-series analysis techniques can be used to analyze the changes in business response status data over time and capture continuous change features of the business response, such as the trend of business response time changes and the fluctuation of response time. The captured features are then converted into real-time business interaction fluency features consistent with the format of business interaction fluency features in the perception anchor set. For example, if the format of business interaction fluency features in the perception anchor set is a sequence composed of the average response time, the maximum response time, and the minimum response time, then the captured features are converted into a sequence in this format.

[0041] Step S124: In the content presentation feature extraction channel, perform content matching degree analysis on the content loading status data, extract the matching features between content presentation and user operation needs, and convert the extracted features into real-time content presentation matching degree features that are consistent with the content presentation matching degree feature format in the perception anchor set.

[0042] In this embodiment, this step is used to extract real-time content presentation fit features. The content presentation feature extraction channel is used to extract content presentation fit features. When performing content matching analysis on content loading status data, content analysis techniques can be used to analyze the degree of matching between the content presented in the content loading status data and the user's operational needs, extracting fit features, such as the matching degree features between content categories and user historical browsing preference categories, and the matching degree features between content titles and user search keywords. The extracted features are then converted into real-time content presentation fit features with the same format as the content presentation fit features in the perception anchor set. For example, if the format of the content presentation fit features in the perception anchor set is a sequence composed of content category matching degree and content title matching degree, then the extracted features are converted into a sequence in this format.

[0043] Step S125: Extract the feature dimension information of the business interaction fluency feature and the content presentation fit feature from the perception anchor point set, and perform dimension calibration on the real-time business interaction fluency feature and the real-time content presentation fit feature respectively. Through dimension calibration, make the two real-time business experience features, the real-time business interaction fluency feature and the real-time content presentation fit feature, consistent with the corresponding feature types in the perception anchor point set, and finally obtain the real-time business experience features.

[0044] In this embodiment, this step is used to perform dimensional calibration on the real-time business experience features to ensure that they are consistent with the corresponding feature types in the perception anchor set. Feature dimensional information is related to the dimensions of the business interaction fluency feature and the content presentation fit feature, such as the number of dimensions and the meaning of each dimension. When calibrating the dimensions of the real-time business interaction fluency feature and the real-time content presentation fit feature, if the number of dimensions of the real-time business interaction fluency feature differs from the number of dimensions of the business interaction fluency feature in the perception anchor set, it is adjusted to make the number of dimensions the same; if the meaning of each dimension of the real-time business interaction fluency feature differs from the meaning of each dimension of the business interaction fluency feature in the perception anchor set, it is adjusted to make the meaning the same. Through dimensional calibration, the two real-time business experience features—the real-time business interaction fluency feature and the real-time content presentation fit feature—are made consistent with the corresponding feature types in the perception anchor set, ultimately yielding the real-time business experience features.

[0045] Step S130: Call the business experience model to perform anchor point matching between the real-time business experience features and the set of perception anchor points, and dynamically adjust the matching weights through the built-in association enhancement module of the model to generate anchor point matching results.

[0046] In this embodiment, this step is used to match real-time service experience features with a set of perception anchor points to obtain anchor point matching results.

[0047] Step S131: Input the real-time business experience features and the set of perception anchors into the feature input module and anchor storage module of the business experience model, respectively. The feature input module performs standardization processing on the real-time business experience features, and the anchor storage module expands the anchor units in the set of perception anchors into a matchable feature matrix.

[0048] In this embodiment, this step is used to input real-time service experience features and a set of perception anchors into the service experience model and perform preprocessing. The service experience model is a model used to match service experience features with the set of perception anchors. The feature input module is the module in the service experience model used to receive and process real-time service experience features, and the anchor storage module is the module in the service experience model used to store the set of perception anchors. When the feature input module standardizes the real-time service experience features, it can use data standardization techniques to convert the real-time service experience features into feature data with a unified standard, for example, converting the values ​​of the real-time service experience features into values ​​within a certain fixed range. When the anchor storage module expands the anchor units in the set of perception anchors into a matchable feature matrix, it arranges the service experience features in each anchor unit in a certain order to form a feature vector, and the feature vectors of all anchor units form the feature matrix.

[0049] Step S132: The initial matching module of the business experience model performs a preliminary comparison between the standardized real-time business experience features and the business experience features corresponding to the anchor units in the feature matrix one by one, calculates the feature similarity between the real-time business experience features and the anchor units participating in the comparison, and obtains an initial similarity value set.

[0050] In this embodiment, this step is used to perform preliminary anchor point matching to obtain an initial similarity value set. The initial matching module is the module in the business experience model used for preliminary anchor point matching. It compares the standardized real-time business experience features with the business experience features corresponding to each anchor unit in the feature matrix one by one, and calculates the feature similarity between the two. The feature similarity can be calculated using similarity calculation techniques, such as cosine similarity calculation techniques, Euclidean distance calculation techniques, etc. By calculating the similarity value between the real-time business experience features and the business experience features corresponding to each anchor unit, all these similarity values ​​constitute the initial similarity value set.

[0051] Step S133: Input the initial similarity value set into the association enhancement module built into the model. The association enhancement module calls the weight adjustment experience in the same business scenario in the historical matching data, and combines it with the auxiliary operation data in the current real-time business interaction data to determine the weight adjustment coefficient.

[0052] In this embodiment, this step is used to determine the weight adjustment coefficient. The association enhancement module is a module in the business experience model used to dynamically adjust the matching weights. Historical matching data is data generated by the business experience model when performing anchor point matching in the past. The weight adjustment experience under the same business scenario is the relevant experience of weight adjustment under the same business scenario in the historical matching data. For example, in a certain business scenario, which anchor point units' matching weights need to be adjusted, and by what magnitude. Auxiliary operation data is the user operation trigger data temporarily stored in step S122, such as continuous trigger interval data of user operations, operation type sequence data, operation termination reason data, etc. The association enhancement module calls the weight adjustment experience under the same business scenario in the historical matching data, combines it with the auxiliary operation data in the current real-time business interaction data, and determines the weight adjustment coefficient through analysis. For example, the size of the weight adjustment coefficient is determined based on the weight adjustment magnitude under the same business scenario in the historical experience and the operation type sequence data in the current auxiliary operation data.

[0053] Step S1331: After receiving the initial similarity value set, the association enhancement module first extracts the current business scenario identifier corresponding to the initial similarity value set, and then retrieves historical matching data under the same business scenario from the model's historical experience database through the scenario identifier.

[0054] In this embodiment, historical matching data under the same business scenario is retrieved. The current business scenario identifier is the identifier of the business scenario corresponding to the initial similarity value set, such as a shopping scenario identifier, an information browsing scenario identifier, etc. The historical experience database is a database used in the business experience model to store historical matching data. Historical matching data under the same business scenario is retrieved from the historical experience database by the scenario identifier, for example, retrieving historical matching data under all shopping scenarios.

[0055] Step S1332: Analyze the retrieved historical matching data, extract the initial similarity value, the adjusted similarity value, and the corresponding business adjustment effect feedback data, calculate the similarity adjustment rules of different anchor units in the business scenario to which the retrieved historical matching data belongs, and obtain the scenario-specific weight adjustment experience value.

[0056] In this embodiment, scenario-specific weight adjustment experience values ​​are obtained. When analyzing the retrieved historical matching data, data analysis techniques can be used to analyze the relationship between the initial similarity value, the adjusted similarity value, and the corresponding business adjustment effect feedback data. For example, analyzing the change relationship between the initial similarity value and the adjusted similarity value, and the relationship between the adjusted similarity value and the business adjustment effect feedback data. By analyzing and calculating the similarity adjustment rules of different anchor units in the business scenario to which the retrieved historical matching data belongs, for example, in a certain business scenario, for a certain anchor unit, when the initial similarity value is within a certain range, what is the range of change of the adjusted similarity value? This yields scenario-specific weight adjustment experience values.

[0057] Step S1333: Extract auxiliary operation data from the current real-time business interaction data. The auxiliary operation data includes the continuous trigger interval of user operations, the sequence of operation types, and the reason for operation termination. Perform feature encoding processing on the extracted auxiliary operation data to obtain operation feature codes.

[0058] In this embodiment, the operation feature code is obtained. The auxiliary operation data is the user operation trigger data in the current real-time business interaction data. When performing feature encoding processing on the extracted auxiliary operation data, feature encoding technology can be used to convert the auxiliary operation data into feature data with a specific encoding format. For example, the continuous trigger interval of user operation can be converted into the corresponding encoded value, the operation type sequence can be converted into the corresponding encoded sequence, and the operation termination reason can be converted into the corresponding encoded value, etc., to obtain the operation feature code.

[0059] Step S1334: Input the scenario-specific weight adjustment experience value and the operation feature encoding into the coefficient calculation submodule of the association enhancement module. The coefficient calculation submodule fuses the two through a multi-feature fusion algorithm to generate the basic weight adjustment coefficient.

[0060] In this embodiment, a basic weight adjustment coefficient is generated. The coefficient calculation submodule is a submodule within the association reinforcement module used to calculate the weight adjustment coefficient. The multi-feature fusion algorithm is an algorithm used to fuse multiple features, such as a weighted average fusion algorithm or a neural network fusion algorithm. The scene-specific weight adjustment empirical value and the operation feature code are input into the coefficient calculation submodule, and the two are fused using the multi-feature fusion algorithm to generate the basic weight adjustment coefficient. For example, the scene-specific weight adjustment empirical value and the operation feature code are assigned different weights, and then a weighted average is calculated to obtain the basic weight adjustment coefficient.

[0061] Step S1335: Obtain the real-time operating load data of the current business, convert the real-time operating load data into a load influence coefficient, correct the basic weight adjustment coefficient through the load influence coefficient, eliminate the interference of business operating load on matching weight, and finally determine the weight adjustment coefficient used to adjust the initial similarity value.

[0062] In this embodiment, the final weight adjustment coefficient is determined. The real-time operational load data of the current business is data related to the current operational load of the business system, such as CPU utilization data and memory utilization data. When converting the real-time operational load data into a load influence coefficient, data conversion technology can be used to convert the real-time operational load data into a coefficient that reflects its influence on the matching weight. For example, when the CPU utilization of the business system is high, the load influence coefficient is large; when the CPU utilization of the business system is low, the load influence coefficient is small. When correcting the basic weight adjustment coefficient using the load influence coefficient, a multiplicative correction method can be used. The basic weight adjustment coefficient is multiplied by the load influence coefficient to obtain the corrected weight adjustment coefficient, eliminating the interference of business operational load on the matching weight, and finally determining the weight adjustment coefficient used to adjust the initial similarity value.

[0063] Step S134: Dynamically adjust the similarity values ​​in the initial similarity value set one by one using the weight adjustment coefficient, increase the similarity value of anchor units that meet the priority matching conditions in terms of relevance to the current business scenario, and decrease the similarity value of anchor units that do not meet the priority matching conditions in terms of relevance, to obtain the adjusted similarity value set.

[0064] In this embodiment, this step is used to dynamically adjust the initial similarity values ​​to obtain an adjusted set of similarity values. For each similarity value in the initial similarity value set, adjustments are made according to the corresponding weight adjustment coefficient, for example, multiplying the similarity value by the weight adjustment coefficient to obtain the adjusted similarity value. Anchor units whose relevance to the current business scenario meets the priority matching criteria are those whose business scenario information is highly relevant to the current business scenario information; for these anchor units, their similarity values ​​are increased. Anchor units whose relevance does not meet the priority matching criteria are those whose business scenario information is less relevant to the current business scenario information; for these anchor units, their similarity values ​​are decreased. Through dynamic adjustment, an adjusted set of similarity values ​​is obtained.

[0065] Step S135: Select anchor units that meet the optimal matching criteria from the adjusted similarity value set, and integrate the identification information of the selected anchor units, the corresponding similarity values, and the business experience features and user perception descriptions contained in the anchor units to generate anchor matching results.

[0066] In this embodiment, this step is used to filter out the optimal matching anchor units and generate anchor matching results. The optimal matching criterion is a pre-set standard for filtering the optimal matching anchor units, such as selecting the anchor unit with the highest adjusted similarity value as the optimal matching anchor unit. After filtering out the anchor units corresponding to the similarity values ​​that meet the optimal matching criterion, the identification information of the anchor unit, the corresponding similarity value, and the business experience features and user perception descriptions contained in the anchor unit are integrated to form the anchor matching results.

[0067] Step S140: Perform perception representation transformation based on the anchor point matching result. Use the user perception representation in the matching anchor point unit corresponding to the real-time business experience feature as the basis, and supplement the detailed description with the real-time business scenario to form real-time user perception information.

[0068] In this embodiment, this step is used to convert the anchor point matching result into real-time user-perceived information, reflecting the user's perception of the current business experience.

[0069] Step S141: Parse the anchor matching result, extract the matching anchor unit, the corresponding real-time business experience features, and the current business scenario information. The current business scenario information is obtained from the scenario identifier field of the real-time business interaction data.

[0070] In this embodiment, this step is used to parse the anchor point matching result and extract relevant information. The anchor point matching result is generated in step S135. When parsing this result, the matching anchor point unit, the corresponding real-time business experience features, and the current business scenario information are extracted. The current business scenario information is obtained from the scenario identifier field of the real-time business interaction data. The scenario identifier field is a field in the real-time business interaction data used to identify the business scenario. For example, a field is set in the real-time business interaction data to store the identification information of the business scenario.

[0071] Step S142: Extract the user perception expression contained in the matching anchor unit as the basic perception expression, associate the basic perception expression with the specific feature performance of the real-time business experience features, and sort out the business experience details corresponding to the basic perception expression.

[0072] In this embodiment, this step is used to extract basic perception statements and sort out the details of the business experience. The user perception statements contained in the matching anchor unit are the user's statements about the business experience corresponding to that anchor unit, and are used as basic perception statements. When associating the basic perception statement with the specific features of the real-time business experience characteristics, the business experience characteristics corresponding to the basic perception statement are analyzed. For example, if the basic perception statement is "very smooth to use," the corresponding business experience characteristic is the smoothness of business interaction. The basic perception statement is then associated with the specific features of the smoothness of real-time business interaction characteristics, sorting out the details of the business experience corresponding to the basic perception statement, such as details like short response time and fast response speed in the smoothness of business interaction characteristics.

[0073] Step S143: Call the scene description dictionary and match the corresponding scene detail description words from the scene description dictionary according to the current business scene information.

[0074] In this embodiment, this step is used to match scenario detail description terms. The scenario description terminology is a database that stores scenario detail description terms corresponding to various business scenarios. When the scenario description terminology is called, the corresponding scenario detail description terms are searched in the scenario description terminology based on the current business scenario information. For example, if the current business scenario is a shopping scenario, scenario detail description terms related to the shopping scenario, such as "product browsing", "shopping cart operation", and "order submission", are searched in the scenario description terminology.

[0075] Step S1431: Construct a scenario description lexicon. The scenario description lexicon adopts a classified storage structure, which is divided into multiple primary categories according to business type. Each primary category is further divided into multiple secondary categories according to specific usage scenarios. Each secondary category corresponds to a set of scenario detail description words.

[0076] In this embodiment, a scenario description thesaurus is constructed. The scenario description thesaurus adopts a categorized storage structure, divided into multiple primary categories according to business type, such as shopping business type, information browsing business type, and social business type; each primary category is further divided into multiple secondary categories according to specific usage scenarios. For example, the specific usage scenarios under the shopping business type can be divided into product search scenarios, product browsing scenarios, shopping cart scenarios, and order submission scenarios; each secondary category corresponds to a set of scenario detail description terms. For example, the scenario detail description terms corresponding to the product search scenario could be "search keyword input" and "search result display," etc.

[0077] Step S1432: Filter the scene detail description words corresponding to the divided secondary categories one by one, select commonly used words that can accurately reflect the scene characteristics of the divided secondary categories, and eliminate rare words and ambiguous words to ensure that the selected words have clear meanings and conform to the user's daily expression habits.

[0078] In this embodiment, vocabulary for describing scene details is used. Each secondary category's vocabulary for describing scene details is filtered one by one, selecting commonly used words that accurately reflect the characteristics of the scene corresponding to that secondary category. For example, in the vocabulary for describing scene details corresponding to the product search scenario, commonly used words such as "search keyword input" and "search result display" are selected. Uncommon and ambiguous words are removed, such as "search" (uncommon word) and "find" (ambiguous, possibly referring to both searching and browsing), ensuring that the selected words are clear in meaning and conform to users' daily expression habits.

[0079] Step S1433: Add semantic tags and scene relevance scores to each scene detail description word in the scene description thesaurus. When the score value meets the preset close matching conditions, the corresponding word is more in line with the requirements of the current business scene.

[0080] In this embodiment, tags and scores are added to scene detail description terms. Semantic tags are labels used to describe the meaning of scene detail description terms. For example, the semantic tags corresponding to "search keyword input" could be "search operation" or "keyword input." Scene relevance score is a score used to measure the degree of relevance between scene detail description terms and their corresponding scenes. The higher the score, the higher the relevance between the term and the scene. When the score value meets a preset close matching condition, the corresponding term is more closely matched with the current business scene. For example, the preset close matching condition is that the scene relevance score is greater than or equal to a certain value. When the scene relevance score of a scene detail description term is greater than or equal to that value, it means that the term is more closely matched with the current business scene.

[0081] Step S1434: Parse the current business scenario information, extract the business type identifier and specific scenario identifier, locate the corresponding first-level category in the scenario description thesaurus based on the business type identifier, and then locate the corresponding second-level category based on the specific scenario identifier.

[0082] In this embodiment, secondary categories are used to locate the scene description thesaurus. When parsing the current business scene information, the business type identifier and the specific scene identifier are extracted. For example, if the current business scene information is "shopping business - product search scene", then the business type identifier is "shopping business" and the specific scene identifier is "product search scene". The corresponding primary category in the scene description thesaurus is located based on the business type identifier. For example, "shopping business" is used to locate the primary category of the shopping business type. Then, the corresponding secondary category is located based on the specific scene identifier. For example, "product search scene" is used to locate the secondary category of the product search scene under the shopping business type.

[0083] Step S1435: In the located secondary categories, filter the scene detail description words that meet the preset standard in scene relevance score, and select the required number of words as matching results according to the order in which the scores meet the preset sorting rules.

[0084] In this embodiment, keywords for filtering and selecting scene detail descriptions are used. The preset standard is a pre-defined scene relevance score threshold, such as a scene relevance score greater than or equal to a certain value; the preset sorting rule is a pre-defined rule for sorting scene detail descriptions, such as sorting them in descending order of scene relevance score; the required quantity is a pre-defined number of scene detail descriptions to be selected, such as selecting a certain number of scene detail descriptions. Within the located secondary categories, scene detail descriptions that meet the preset standard in scene relevance score are filtered out, sorted according to the preset sorting rule, and keywords meeting the required quantity are selected as the matching results.

[0085] Step S144: Integrate the scene detail description vocabulary into the basic perception expression, and expand the basic perception expression by combining the specific manifestations of real-time business experience characteristics, so that the expanded expression not only retains the core of the user's direct experience, but also includes real-time business scenarios and experience details.

[0086] In this embodiment, this step is used to expand the basic perception description to form a perception description that includes scene details and experience details. Scene detail descriptions are integrated into the basic perception description. For example, if the basic perception description is "very smooth to use," and the scene detail description is "product search," then the integration results in "very smooth to use during product search." The basic perception description is further expanded by combining the specific manifestations of real-time business experience features. For example, if the specific manifestation of real-time business experience features is "short response time," then the expansion is "very smooth to use during product search, with a short response time." Through this expansion, the expanded description retains the core user experience while also including real-time business scenarios and experience details.

[0087] Step S145: Perform semantic fluency checks on the expanded perception representation, adjust the sentence structure to conform to natural language expression habits, remove redundant words, and finally form real-time user perception information.

[0088] In this embodiment, this step is used to optimize the expanded perceptual representation to form the final real-time user perception information. Semantic fluency checking examines whether the expanded perceptual representation is semantically fluent, such as checking for grammatical errors and appropriate vocabulary usage. The sentence structure is adjusted to conform to natural language expression habits; for example, "It's very smooth to use, during product search" is changed to "It's very smooth to use during product search." Redundant words are removed, such as repeated or unnecessary words in the representation. Through optimization, the final real-time user perception information is formed.

[0089] Step S150: Compare the real-time user perception information with the preset perception optimization benchmark to generate perception optimization guidelines that include perception improvement directions and specific business adjustment suggestions.

[0090] In this embodiment, this step is used to generate perception optimization guidance based on real-time user perception information, providing direction and suggestions for business optimization.

[0091] Step S151: Obtain a preset perception optimization benchmark. The perception optimization benchmark includes the perception expression standard of user satisfaction under different business scenarios and the corresponding business experience feature threshold. The perception expression standard is divided into multiple levels, and each level corresponds to a set of business experience feature thresholds.

[0092] In this embodiment, this step is used to obtain the perception optimization benchmark. The perception optimization benchmark is a pre-set benchmark used to measure whether user perception meets the satisfaction standard. It includes the perception expression standard of user satisfaction under different business scenarios and the corresponding business experience feature thresholds. The perception expression standard is divided into multiple levels, such as five levels: very satisfied, satisfied, average, dissatisfied, and very dissatisfied. Each level corresponds to a set of business experience feature thresholds. For example, the business interaction smoothness feature threshold corresponding to the very satisfied level is a response time less than a certain value, and the content presentation fit feature threshold is a matching degree greater than a certain value.

[0093] Step S152: Compare the real-time user perception information with the perception description standard in the perception optimization benchmark to determine the current perception level to which the real-time user perception information belongs, and extract the business experience feature threshold range corresponding to the current perception level.

[0094] In this embodiment, this step is used to determine the perception level to which the real-time user perception information belongs and extract the corresponding business experience feature threshold range. The real-time user perception information is compared with each perception description standard level in the perception optimization benchmark. For example, if the real-time user perception information is "the product search process is very smooth, and the response time is short," it is compared with the satisfaction level perception description standard in the perception optimization benchmark. If it meets the satisfaction level perception description standard, then the current perception level to which the real-time user perception information belongs is determined to be the satisfaction level. Simultaneously, the business experience feature threshold range corresponding to this current perception level is extracted. For example, the business interaction smoothness feature threshold range corresponding to the satisfaction level is a response time within a certain numerical range, and the content presentation fit feature threshold range is a matching degree within a certain numerical range, etc.

[0095] Step S153: Compare the real-time service experience features with the service experience feature threshold range corresponding to the current perception level, and find the feature items in the real-time service experience features that exceed the threshold range. The feature items that exceed the threshold range are the key factors that cause the current perception level to fail to meet the preset ideal standard.

[0096] In this embodiment, this step is used to identify the key factors causing the current perception level to fall short of the ideal standard. Each feature item in the real-time service experience features is compared with the threshold range of the service experience features corresponding to the current perception level. For example, whether the value of the response time feature item in the real-time service interaction fluency feature is within the corresponding threshold range, or whether the value of the matching degree feature item in the content presentation fit feature is within the corresponding threshold range, etc. If the value of a certain feature item exceeds the corresponding threshold range, then that feature item is the key factor causing the current perception level to fall short of the preset ideal standard.

[0097] Step S154: For the identified key factors, call the business optimization suggestion library one by one to match specific business adjustment suggestions that can improve the corresponding key factors.

[0098] In this embodiment, this step is used to match business adjustment suggestions that can improve key factors.

[0099] For example, in step S1541: a business optimization suggestion library is constructed. The business optimization suggestion library adopts a key-value pair storage method, with key factors as keys and corresponding business adjustment suggestions as values. Each of the key factors included corresponds to at least one set of business adjustment suggestions.

[0100] In this embodiment, a business optimization suggestion library is constructed. This library is a database storing key factors and corresponding business adjustment suggestions, using a key-value pair storage method. The key is the key factor, and the value is the corresponding business adjustment suggestion. Each included key factor corresponds to at least one set of business adjustment suggestions. For example, if the key factor is "too long response time," the corresponding business adjustment suggestions could be "optimize server performance" or "reduce page resource loading."

[0101] Step S1542: The collected business adjustment suggestions are structured one by one, and the resulting business adjustment suggestion groups all include the parameter adjustment direction, specific operation method, operation step sequence, and operation precautions. The parameter adjustment direction, specific operation method, operation step sequence, and operation precautions are all expressed in the form of clear operation instructions.

[0102] In this embodiment, business adjustment suggestions are structured. Each suggestion is broken down into parts such as parameter adjustment direction, specific operation method, operation step sequence, and operation precautions. For example, if the business adjustment suggestion is "optimize server performance," the parameter adjustment direction could be "increase the server CPU utilization limit," the specific operation method could be "modify the CPU utilization limit parameter in the server configuration file," the operation step sequence could be "open the server configuration file, find the CPU utilization limit parameter, modify the parameter value, save the configuration file, and restart the server," and the operation precautions could be "when modifying the parameter value, it should be set reasonably according to the server hardware configuration to avoid setting it too high, which may cause server instability." The parameter adjustment direction, specific operation method, operation step sequence, and operation precautions are all expressed in the form of clear operation instructions to ensure that business personnel can operate according to the instructions.

[0103] Step S1543: Add feature tags to each of the key factors included in the business optimization suggestion library, and the feature tags correspond one-to-one with the feature items in the real-time business experience features.

[0104] In this embodiment, feature tags are added to key factors. Feature tags are labels used to identify the real-time business experience feature items corresponding to the key factors. For example, if the key factor is "excessive response time," and the corresponding real-time business experience feature item is the response time feature item in the business interaction fluency feature, then the feature tag added to this key factor is "business interaction fluency - response time." The feature tags correspond one-to-one with the feature items in the real-time business experience feature, ensuring accurate matching between key factors and their corresponding real-time business experience feature items.

[0105] Step S1544: For the identified key factors, extract the feature tags of the corresponding key factors one by one, use the feature tags as search keywords to search in the business optimization suggestion library, and locate the key factor entries that completely match the feature tags.

[0106] In this embodiment, key factor entries are retrieved from the business optimization suggestion library. For each identified key factor, its corresponding feature tag is extracted. For example, if the key factor is "excessive response time," the extracted feature tag is "business interaction smoothness - response time." This feature tag is used as a search keyword to search the business optimization suggestion library for key factor entries that perfectly match the feature tag. For example, the key factor entry with the key "business interaction smoothness - response time" is searched in the business optimization suggestion library.

[0107] Step S1545: Extract the corresponding business adjustment suggestions from the identified key factor items, and fine-tune the operation parameters in the business adjustment suggestions in combination with the real-time operation status of the current business to make the adjustment suggestions more in line with the actual situation of the current business.

[0108] In this embodiment, business adjustment suggestions are extracted and fine-tuned. Corresponding business adjustment suggestions are extracted from the identified key factor entries. For example, from the key factor entry with the key "business interaction smoothness - response time", the corresponding business adjustment suggestion "optimize server performance" is extracted. The operation parameters in the business adjustment suggestions are fine-tuned based on the real-time operating status of the current business. For example, if the server CPU utilization rate of the current business is already high, the operation parameters in the business adjustment suggestion "optimize server performance" are fine-tuned, changing "increase the upper limit of server CPU utilization" to "rationally allocate server CPU resources, prioritizing CPU usage for business interaction-related processes," making the adjustment suggestions more aligned with the actual situation of the current business.

[0109] Step S1546: Verify the feasibility of the fine-tuned business adjustment suggestions, confirm that the operation methods in the suggestions are executable in the current business environment, and take the verified business adjustment suggestions as specific business adjustment suggestions that can improve the corresponding key factors.

[0110] In this embodiment, the feasibility of business adjustment suggestions is verified, and the final specific business adjustment suggestions are determined. When verifying the feasibility of the fine-tuned business adjustment suggestions, it is analyzed whether the operational methods in the suggestions are executable in the current business environment. For example, it is analyzed whether the operational method of "reasonably allocating server CPU resources and prioritizing CPU usage for business interaction-related processes" has the corresponding technical conditions and resource support in the current business environment. If the verification passes, the business adjustment suggestion is adopted as a specific business adjustment suggestion that can improve the corresponding key factors; if the verification fails, the business adjustment suggestion is fine-tuned again until the verification passes.

[0111] Step S155: Determine the direction of perception improvement based on the gap between the current perception level and the optimal perception level, link and integrate the perception improvement direction with the corresponding specific business adjustment suggestions, supplement the description of the expected improvement effect of each adjustment suggestion, and form a perception optimization guide.

[0112] In this embodiment, this step is used to determine the direction of perception improvement and generate perception optimization guidelines. The optimal perception level is the highest level in the perception optimization benchmark, such as the "very satisfactory" level. The perception improvement direction is determined based on the gap between the current perception level and the optimal perception level. For example, if the current perception level is the "satisfactory" level and the optimal perception level is the "very satisfactory" level, then the perception improvement direction is to improve from the "satisfactory" level to the "very satisfactory" level. The perception improvement direction is then associated and integrated with the corresponding specific business adjustment suggestions. For example, the perception improvement direction of "improving from the satisfactory level to the "very satisfactory" level is associated and integrated with specific business adjustment suggestions such as "optimizing server performance" and "reducing page resource loading." The expected improvement effect descriptions of each adjustment suggestion are supplemented, such as "optimizing server performance is expected to shorten the response time by a certain percentage" and "reducing page resource loading is expected to shorten the page loading time by a certain percentage." Through association, integration, and supplementation of expected improvement effect descriptions, perception optimization guidelines are formed.

[0113] Figure 2 The illustration shows exemplary hardware and software components of a user-aware mapping system 100 based on a business experience model, which can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the user-aware mapping system 100 based on a business experience model and to perform the functions in this application.

[0114] The user-aware mapping system 100 based on the business experience model can be a general-purpose server or a special-purpose server; both can be used to implement the user-aware mapping method based on the business experience model of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0115] For example, a user-aware mapping system 100 based on a service experience model may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the user-aware mapping system 100 based on a service experience model may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The user-aware mapping system 100 based on a service experience model also includes an I / O interface 150 between the computer and other input / output devices.

[0116] For ease of explanation, only one processor is described in the user-aware mapping system 100 based on the business experience model. However, it should be noted that the user-aware mapping system 100 based on the business experience model in this application may also include multiple processors. Therefore, the steps executed by one processor as described in this application may also be executed jointly by multiple processors or individually. For example, if the processor of the user-aware mapping system 100 based on the business experience model executes steps A and B, it should be understood that steps A and B may also be executed jointly by two different processors or individually by one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor jointly execute steps A and B.

[0117] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the user perception mapping method based on the business experience model as described above is implemented.

[0118] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A user perception mapping method based on a business experience model, characterized in that, The method includes: Construct a set of perception anchors for business experience features and user perception expressions. The business experience features include the smoothness of business interaction and the fit of content presentation. The user perception expressions include the direct feelings and descriptions of users when using the business. The real-time business interaction data is processed by feature parsing to extract real-time business experience features that are consistent with the business experience feature types in the set of perception anchors. The real-time business interaction data is dynamic data generated by users during the use of the business. The business experience model is invoked to match the real-time business experience features with the set of perception anchors. The matching weights are dynamically adjusted through the association enhancement module built into the model to generate anchor matching results. Based on the anchor point matching results, a perception representation transformation is performed. The user perception representation in the matching anchor point unit corresponding to the real-time business experience features is used as the basis, and detailed descriptions are added in combination with the real-time business scenario to form real-time user perception information. The real-time user perception information is compared with a preset perception optimization benchmark to generate perception optimization guidelines that include perception improvement directions and specific business adjustment suggestions.

2. The user perception mapping method based on a business experience model according to claim 1, characterized in that, The set of perception anchors for constructing business experience features and user perception expressions includes the smoothness of business interaction and the fit of content presentation. The user perception expressions include direct descriptions of users' feelings about using the business, including: Collect historical interaction datasets and synchronized user feedback datasets during business usage; Business interaction-related data and content presentation-related data are extracted from the historical interaction dataset. The business interaction-related data is subjected to fluency feature quantification to obtain business interaction fluency features. The content presentation-related data is subjected to fit feature extraction to obtain content presentation fit features. Semantic segmentation is performed on the user subjective feelings text in the synchronous user feedback dataset to extract the core sentences that can directly reflect the user's experience as the user perception expression, and text content that is vague or irrelevant to the business experience is removed. The characteristics of smooth business interaction and content presentation fit are associated and bound with the corresponding user perception expression to form an initial anchor unit. Each initial anchor unit formed by association and binding contains a set of characteristics of smooth business interaction, content presentation fit and user perception expression. Duplicate items are removed from all initial anchor units. For each initial anchor unit, the correlation stability between business experience features and user perception statements is calculated sequentially. Initial anchor units with correlation stability that meet the preset standard are retained. The retained initial anchor units are classified and organized according to business scenario type to form a set of perception anchors for business experience features and user perception statements.

3. The user perception mapping method based on a business experience model according to claim 1, characterized in that, The process involves feature parsing of real-time business interaction data to extract real-time business experience features that match the business experience feature types in the perception anchor set. The real-time business interaction data refers to dynamic data generated during user interaction with the service, including: The real-time business interaction data generated by users during the use of services is obtained through the business data collection interface. The real-time business interaction data includes business response status data, content loading status data, and user operation trigger data. The real-time business interaction data is processed by data diversion, separating the business response status data and content loading status data, and transmitting them to the corresponding feature extraction channels respectively. User operation trigger data is temporarily stored as auxiliary data. In the business response feature extraction channel, time series features are extracted from the business response status data to capture the continuous change features of the business response. The captured features are then converted into real-time business interaction smoothness features that are consistent with the business interaction smoothness feature format in the perception anchor set. In the content presentation feature extraction channel, content matching degree analysis is performed on the content loading status data to extract the matching features between content presentation and user operation needs. The extracted features are then converted into real-time content presentation matching degree features that are consistent with the content presentation matching degree feature format in the perception anchor set. Extract the feature dimension information of the business interaction fluency feature and the content presentation fit feature from the perception anchor point set, and perform dimension calibration on the real-time business interaction fluency feature and the real-time content presentation fit feature respectively. Through dimension calibration, make the two real-time business experience features, the real-time business interaction fluency feature and the real-time content presentation fit feature, consistent with the corresponding feature types in the perception anchor point set, and finally obtain the real-time business experience features.

4. The user perception mapping method based on a business experience model according to claim 1, characterized in that, The invoked business experience model performs anchor point matching between the real-time business experience features and the set of perception anchor points. The matching weights are dynamically adjusted through the model's built-in association enhancement module to generate anchor point matching results, including: The real-time business experience features and the set of perception anchors are respectively input into the feature input module and the anchor storage module of the business experience model. The feature input module performs standardization processing on the real-time business experience features, and the anchor storage module expands the anchor units in the set of perception anchors into a matchable feature matrix. The initial matching module of the business experience model performs a preliminary comparison between the standardized real-time business experience features and the business experience features corresponding to the anchor units in the feature matrix one by one, calculates the feature similarity between the real-time business experience features and the anchor units participating in the comparison, and obtains an initial similarity value set. The initial similarity value set is input into the association enhancement module built into the model. The association enhancement module calls the weight adjustment experience in the same business scenario in the historical matching data, and combines it with the auxiliary operation data in the current real-time business interaction data to determine the weight adjustment coefficient. The similarity values ​​in the initial similarity value set are dynamically adjusted one by one by the weight adjustment coefficients. The similarity values ​​of anchor units that meet the priority matching conditions in terms of relevance to the current business scenario are increased, while the similarity values ​​of anchor units that do not meet the priority matching conditions are decreased, resulting in the adjusted similarity value set. From the adjusted set of similarity values, anchor units corresponding to similarity values ​​that meet the optimal matching criteria are selected. The identification information of the selected anchor units, their corresponding similarity values, and the business experience features and user perception descriptions contained in the anchor units are integrated to generate anchor matching results.

5. The user perception mapping method based on a business experience model according to claim 1, characterized in that, The process of performing perception representation transformation based on the anchor point matching result involves using the user perception representation in the matching anchor point unit corresponding to the real-time business experience features as a basis, supplementing it with detailed descriptions based on the real-time business scenario, and forming real-time user perception information, including: The anchor matching result is parsed, and the matching anchor unit, the corresponding real-time business experience feature, and the current business scenario information are extracted. The current business scenario information is obtained from the scenario identifier field of the real-time business interaction data. Extract the user perception statements contained in the matching anchor unit as basic perception statements, associate the basic perception statements with the specific feature manifestations of real-time business experience features, and sort out the business experience details corresponding to the basic perception statements. The scenario description dictionary is invoked, and corresponding scenario detail description words are matched from the scenario description dictionary based on the current business scenario information; The scenario detail descriptions are integrated into the basic perception descriptions, and the basic perception descriptions are expanded by combining the specific manifestations of real-time business experience features. This expands the descriptions so that they retain the core of the user's direct experience while also including real-time business scenarios and experience details. The expanded perception statement is subjected to semantic fluency checks, the sentence structure is adjusted to conform to natural language expression habits, redundant words are removed, and finally real-time user perception information is formed.

6. The user perception mapping method based on a business experience model according to claim 1, characterized in that, The step of comparing the real-time user perception information with a preset perception optimization benchmark to generate a perception optimization guide that includes perception improvement directions and specific business adjustment suggestions includes: Obtain a preset perception optimization benchmark, which includes the perception expression standard of user satisfaction under different business scenarios and the corresponding business experience feature threshold. The perception expression standard is divided into multiple levels, and each level corresponds to a set of business experience feature thresholds. The real-time user perception information is compared hierarchically with the perception description standard in the perception optimization benchmark to determine the current perception level to which the real-time user perception information belongs, and the threshold range of business experience features corresponding to the current perception level is extracted. The real-time service experience features are compared with the service experience feature threshold range corresponding to the current perception level. Feature items that exceed the threshold range in the real-time service experience features are identified. Feature items that exceed the threshold range are the key factors that cause the current perception level to fail to meet the preset ideal standard. For each identified key factor, the business optimization suggestion library is invoked one by one to match specific business adjustment suggestions that can improve the corresponding key factor. Based on the gap between the current perception level and the optimal perception level, determine the direction for perception improvement, link and integrate the perception improvement direction with the corresponding specific business adjustment suggestions, supplement the description of the expected improvement effect of each adjustment suggestion, and form a perception optimization guide.

7. The user perception mapping method based on the business experience model according to claim 4, characterized in that, The initial similarity value set is input into the association enhancement module built into the model. The association enhancement module calls the weight adjustment experience in the same business scenario in the historical matching data, and combines it with the auxiliary operation data in the current real-time business interaction data to determine the weight adjustment coefficient, including: After receiving the initial similarity value set, the association enhancement module first extracts the current business scenario identifier corresponding to the initial similarity value set, and then retrieves historical matching data under the same business scenario from the model's historical experience database using the scenario identifier. Analyze the retrieved historical matching data, extract the initial similarity value, the adjusted similarity value, and the corresponding business adjustment effect feedback data, calculate the similarity adjustment rules of different anchor units in the business scenario to which the retrieved historical matching data belongs, and obtain scenario-specific weight adjustment experience values. Auxiliary operation data is extracted from the current real-time business interaction data. The auxiliary operation data includes the continuous trigger interval of user operations, the sequence of operation types, and the reason for operation termination. The extracted auxiliary operation data is subjected to feature encoding processing to obtain operation feature codes. The scenario-specific weight adjustment experience value and the operation feature encoding are input into the coefficient calculation submodule of the association enhancement module. The coefficient calculation submodule fuses the two through a multi-feature fusion algorithm to generate the basic weight adjustment coefficient. Obtain real-time operational load data of the current business, convert the real-time operational load data into load impact coefficients, and use the load impact coefficients to correct the basic weight adjustment coefficients, thereby eliminating the interference of business operational load on matching weights, and finally determining the weight adjustment coefficients used to adjust the initial similarity values.

8. The user perception mapping method based on the business experience model according to claim 5, characterized in that, The process of calling the scenario description dictionary involves matching corresponding scenario detail description terms from the scenario description dictionary based on the current business scenario information, including: A scenario description thesaurus is constructed. The scenario description thesaurus adopts a classified storage structure, which is divided into multiple primary categories according to business type. Each primary category is further divided into multiple secondary categories according to specific usage scenarios. Each secondary category corresponds to a set of scenario detail description words. The vocabulary describing the scene details corresponding to the division of the secondary categories was screened one by one. Commonly used words that can accurately reflect the scene characteristics of the division of the secondary categories were selected, and obscure and ambiguous words were eliminated to ensure that the selected words are clear in meaning and conform to the user's daily expression habits. Semantic tags and scenario relevance scores are added to each scenario detail description word in the scenario description thesaurus. When the score value meets the preset close matching conditions, the corresponding word is more relevant to the current business scenario. The current business scenario information is parsed, and the business type identifier and specific scenario identifier are extracted. The corresponding first-level category in the scenario description thesaurus is located based on the business type identifier, and the corresponding second-level category is located based on the specific scenario identifier. Within the identified secondary categories, filter out scene detail descriptions that meet the preset criteria for scene relevance scores, and select the required number of words as matching results according to the order in which the scores conform to the preset sorting rules.

9. The user perception mapping method based on the business experience model according to claim 2, characterized in that, The process involves removing duplicates from all initial anchor units, calculating the correlation stability between business experience features and user perception expressions for each initial anchor unit, retaining initial anchor units whose correlation stability meets preset standards, and classifying and organizing the retained initial anchor units according to business scenario types to form a set of perception anchors for business experience features and user perception expressions, including: All initial anchor units are subjected to feature hashing, and a unique feature hash value is generated for each initial anchor unit. Duplicate initial anchor units are identified by comparing feature hash values. One duplicate unit is retained and the others are deleted to complete the duplicate item removal process. A unique identifier is assigned to each of the initial anchor units after duplicates are removed. The business experience feature data sequence and the corresponding user perception expression semantic vector are extracted from the processed initial anchor units and input into the association stability calculation module. The association stability calculation module calculates the degree of association between the business experience feature data sequence and the user perception expression semantic vector through the semantic association analysis algorithm, and generates an association stability value. When the association stability value reaches the preset stability standard, the association between the business experience feature data sequence and the user perception expression semantic vector meets the requirements of actual application. Set a preset standard for correlation stability, compare the correlation stability values ​​of the processed initial anchor units one by one with the preset standard, retain the initial anchor units whose correlation stability values ​​reach or exceed the preset standard, and remove the initial anchor units whose values ​​are lower than the preset standard. Extract the business scenario information corresponding to the retained initial anchor units, classify the retained initial anchor units according to the business scenario information, and group the initial anchor units belonging to the same business scenario into one category to form a scenario anchor group. Each initial anchor unit in the divided scenario anchor group undergoes feature consistency verification to ensure that the business experience feature type of the initial anchor unit in the same group is consistent. All scenario anchor groups are then integrated to form a set of perception anchors that represent business experience features and user perception expressions.

10. A user perception mapping system based on a business experience model, characterized in that, The user perception mapping system based on the business experience model includes a processor and a memory, the memory and the processor are connected, the memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the user perception mapping method based on the business experience model as described in any one of claims 1-9.

Images