Enterprise intelligent management system and method based on cloud platform

Abstract

The invention discloses an enterprise intelligent management system and method based on a cloud platform, and the method comprises the steps: extracting a use feedback questionnaire of a target customer from a questionnaire survey storage module, and extracting use feedback text content from the use feedback questionnaire; and then segmenting and semantic coding of the use feedback text content are carried out by adopting a natural language processing technology and a coding mode based on deep learning, so that condition description semantic enhancement is carried out on each coded question-reply semantic coding feature, and intelligent judgment of an emotion polarity label of each question level is realized. And determining whether the target customer is satisfied with the product. Compared with a simple scoring method, the method has the advantages that fine-grained analysis can be performed on the use feedback of the customer, so that the complex emotion of the customer can be captured, the actual feeling and specific problems of the customer can be understood, the emotion polarity of the customer can be accurately judged, and comprehensive and accurate analysis on the customer feedback can be realized.

Description

Technical Field

[0001] The present application relates to the field of intelligent management, and more specifically, to an enterprise intelligent management system and method based on a cloud platform. Background Art

[0002] Enterprise intelligence management systems rely on cloud technology to improve enterprise operational efficiency, service quality and customer satisfaction. Currently, such systems are insufficient in understanding customer and employee loyalty and satisfaction, which may not only cause customer and employee churn, but also lead to performance decline and decision-making errors.

[0003] In this regard, patent CN118627733A proposes an enterprise intelligent management system and method, which collects and analyzes customer data to evaluate customer loyalty and satisfaction, then collects enterprise operation data, analyzes market share and product competitiveness, and evaluates employee loyalty and satisfaction. Then, it uses the above data for comprehensive analysis and risk prediction, and triggers corresponding risk warnings based on the prediction results. Finally, based on the warning information and combined with historical cases, it formulates emergency strategies to prevent customer and employee loss, reduce business risks, and ensure the healthy and sustainable development of the enterprise.

[0004] The customer satisfaction in the above patent is based on the ratio of the customer's score in the questionnaire content to the preset value. However, the scoring-based method can only capture the customer's superficial satisfaction with the product or service, and cannot deeply understand the customer's actual feelings and specific problems. For example, a customer may give a high score to a certain aspect, but still be dissatisfied during actual use. These subtle dissatisfaction and specific problems cannot be reflected by a simple score. In addition, the scoring method cannot capture the customer's complex emotions, such as contradictions, hesitations, or emotional changes in specific situations. For example, a customer may be satisfied with the product function, but disappointed with the after-sales service. This emotional contradiction and emotional changes in specific situations cannot be fully reflected by scoring.

[0005] Therefore, an optimized enterprise intelligent management solution is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an enterprise intelligent management system and method based on a cloud platform, which extracts the target customer's usage feedback questionnaire from the questionnaire storage module and extracts the usage feedback text content, and then uses natural language processing technology and encoding methods based on deep learning to perform segmentation and semantic encoding of the usage feedback text content, thereby conditionally describing and enhancing the semantic encoding features of each encoded question-response semantic coding feature, realizing intelligent judgment of the emotional polarity label of each question level, and determining whether the target customer is satisfied with the product. Compared with a simple scoring method, in this way, a fine-grained analysis of customer usage feedback can be performed to capture the customer's complex emotions, understand the customer's actual feelings and specific problems, thereby accurately judging the customer's emotional polarity and achieving a comprehensive and accurate analysis of customer feedback.

[0007] According to one aspect of the present application, there is provided an enterprise intelligent management system based on a cloud platform, which includes: A questionnaire storage module is used to store the collected customer usage feedback questionnaires; a customer management module is used to analyze customer satisfaction with the product based on the customer usage feedback questionnaires; an enterprise management module is used to monitor the enterprise's market share and competitive product dynamics, and evaluate employees' sense of identification with the enterprise; a risk analysis module is used to analyze and predict the risks of the enterprise's operations based on customer satisfaction with the product and employees' sense of identification with the enterprise, and obtain analysis results; an early warning processing module is used to generate corresponding early warnings and formulate corresponding emergency strategies based on the analysis results.

[0008] According to another aspect of the present application, a cloud platform-based enterprise intelligent management method is provided, which includes: Store the collected customer feedback questionnaires; Analyze customer satisfaction with the product based on the customer feedback questionnaire; Monitor the company's market share and competitive dynamics, and assess employees' identification with the company; Based on customer satisfaction with the product and employees' identification with the enterprise, analyze and predict the risks of the enterprise's operations and obtain analysis results; Based on the analysis results, corresponding warnings are generated and corresponding emergency strategies are formulated.

[0009] Compared with the prior art, the present application provides an enterprise intelligent management system and method based on a cloud platform, which extracts the target customer's usage feedback questionnaire from the questionnaire storage module and extracts the usage feedback text content, and then uses natural language processing technology and encoding methods based on deep learning to segment and semantically encode the usage feedback text content, thereby conditionally describing and reinforcing the semantic encoding features of each encoded question-response semantic code to achieve intelligent judgment of the emotional polarity label of each question level and determine whether the target customer is satisfied with the product. Compared with a simple scoring method, this method can perform a fine-grained analysis of customer usage feedback to capture the customer's complex emotions, understand the customer's actual feelings and specific problems, and accurately judge the customer's emotional polarity, and achieve a comprehensive and accurate analysis of customer feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 A block diagram of an enterprise intelligent management system based on a cloud platform according to an embodiment of the present application; Figure 2 A data flow diagram of a cloud platform-based enterprise intelligent management system according to an embodiment of the present application; Figure 3 A block diagram of a customer management module in a cloud platform-based enterprise intelligent management system according to an embodiment of the present application; Figure 4 The present invention is a flowchart of a cloud platform-based enterprise intelligent management method according to an embodiment of the present application. DETAILED DESCRIPTION

[0012] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0013] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0014] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0015] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0016] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0017] In the technical solution of the present application, a cloud platform-based enterprise intelligent management system is proposed. Figure 1 The block diagram is a cloud platform-based enterprise intelligent management system according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of cloud-based enterprise intelligent management system according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the cloud platform-based enterprise intelligent management system 300 of the embodiment of the present application, it includes: a questionnaire storage module 310, which is used to store the collected customer usage feedback questionnaires; a customer management module 320, which is used to analyze the customer's satisfaction with the product based on the customer's usage feedback questionnaire; an enterprise management module 330, which is used to monitor the enterprise's market share and competitive product dynamics, and evaluate employees' sense of identification with the enterprise; a risk analysis module 340, which is used to analyze and predict the risks of the enterprise's operations based on customer satisfaction with the product and employee identification with the enterprise, and obtain analysis results; an early warning processing module 350, which is used to generate corresponding early warnings and formulate corresponding emergency strategies based on the analysis results.

[0018] In particular, the questionnaire storage module 310 is used to store the collected customer usage feedback questionnaires. In the enterprise intelligent management system, storing the collected customer usage feedback questionnaires is a key task, which directly affects the accuracy of subsequent data analysis and customer satisfaction evaluation.

[0019] First of all, the design and distribution of the questionnaire is the starting point of the whole process. When designing the questionnaire, it is necessary to ensure that the content of the questionnaire can fully cover all aspects of the customer's product or service, including but not limited to price, performance, quality and service. The questionnaire should be as concise and clear as possible to avoid customers being confused or bored during the filling process. The questionnaire can include multiple-choice questions, scoring questions and open-ended questions to collect detailed feedback from customers. Multiple-choice questions and scoring questions can provide quantitative data for subsequent statistical analysis; open-ended questions can provide detailed descriptions and specific opinions of customers, helping companies to have a deeper understanding of customers' actual feelings and specific problems. After the questionnaire is designed, it is distributed to customers through various channels such as email, social media, and website pop-ups to ensure that the questionnaire distribution channels are diverse to cover as many target customer groups as possible. For example, the questionnaire link can be sent to customers through the company's official website, mobile application, social media platform (such as WeChat, Weibo), email list, etc. In addition, customers can be actively invited to participate in the questionnaire survey during the customer service process to increase the recovery rate of the questionnaire.

[0020] After the customer fills in the questionnaire and submits it, the system receives the questionnaire data submitted by the customer, which is usually stored in a database in electronic form. In order to ensure the integrity of the data, the system should have an automatic verification function to check whether the submitted data meets the expected format and requirements. If the data is found to be abnormal or missing, the customer should be notified in time to re-fill or supplement the relevant information. After the customer submits the questionnaire, the system should provide confirmation information to inform the customer that the questionnaire has been successfully submitted and thank the customer for participating. The confirmation information can be sent through page prompts, pop-ups or emails to ensure that customers can receive feedback in a timely manner. In addition, a small gift or coupon can be provided as a reward after the questionnaire is submitted to encourage customers to actively participate in the questionnaire survey.

[0021] Next, the collected data is initially cleaned to remove invalid or incomplete questionnaire records to ensure the integrity and accuracy of the data. The steps of data cleaning include removing questionnaire records that are not completed or the content is obviously illogical, standardizing data, converting data in different formats into a unified standard format to ensure data consistency, and checking and deleting duplicate questionnaire records to avoid data redundancy.

[0022] The collected customer feedback questionnaires are stored in the questionnaire storage module, which should have efficient data storage and retrieval functions and be able to quickly access a large amount of questionnaire data. The design of the storage module should take into account the data structure and define the appropriate data structure to ensure that the questionnaire data can be efficiently stored and retrieved. For example, a relational database (such as MySQL, PostgreSQL) or a NoSQL database (such as MongoDB) can be used to store questionnaire data. In addition, indexes are created for commonly used query fields to increase data retrieval speed, and data sharding technology is used to store data on multiple nodes to improve storage and query performance. Ensure the security of data storage and take necessary encryption measures to prevent data leakage or unauthorized access. Data security measures include encrypted storage of sensitive data to ensure that the data cannot be directly read even if it is illegally accessed, set up a strict permission management mechanism to ensure that only authorized personnel can access and manage questionnaire data, and record logs of data access and operations to track and audit when security incidents occur.

[0023] In particular, the customer management module 320 is used to analyze the customer's satisfaction with the product based on the customer's usage feedback questionnaire. Figure 3 As shown, the customer management module 320 includes: a questionnaire extraction unit 321, which is used to extract the target customer's usage feedback questionnaire from the questionnaire storage module; a feedback text content extraction unit 322, which is used to extract the usage feedback text content from the target customer's usage feedback questionnaire; a question-reply semantic encoding unit 323, which is used to perform content segmentation and semantic encoding on the usage feedback text content to obtain a set of question-reply semantic features; a satisfaction analysis unit 324, which is used to perform sequence condition description distinguishable reinforcement on the set of question-reply semantic features to obtain a set of question-reply reinforcement semantic features, and determine whether the target customer is satisfied with the product based on the set of question-reply reinforcement semantic features.

[0024] Specifically, the questionnaire extraction unit 321 is used to extract the target customer's usage feedback questionnaire from the questionnaire storage module. It should be understood that through the questionnaire survey, the enterprise can collect a large amount of customer feedback data, which covers the customer's true feelings and specific opinions on various aspects of the product or service. When the customer mentions a specific usage problem or suggestion in the questionnaire, the enterprise can quickly locate the problem through these feedback information and take corresponding measures to improve it. This can not only improve customer satisfaction, but also help enterprises continuously optimize products and services. In addition, extracting the target customer's usage feedback questionnaire helps enterprises provide more personalized services. By conducting in-depth analysis of the feedback from a specific customer, the enterprise can better understand the specific needs and pain points of the customer, thereby providing more customized products and services. Therefore, in the technical solution of the present application, first, the target customer's usage feedback questionnaire is extracted from the questionnaire storage module.

[0025] Specifically, the feedback text content extraction unit 322 is used to extract the usage feedback text content from the usage feedback questionnaire of the target customer. It should be understood that although the multiple-choice questions and scoring questions in the questionnaire can provide some basic satisfaction data, they often cannot fully reflect the actual feelings and specific problems of the customers. The open-ended text feedback content contains the detailed description and specific opinions of the customers, which can provide deeper information. In addition, the text feedback content usually contains specific problems and suggestions encountered by customers in the process of using products or services. These problems may be technical or on the service process. By extracting and analyzing these text contents, enterprises can quickly identify specific problem points and take targeted improvement measures. By extracting the usage feedback text content from the usage feedback questionnaire of the target customer, enterprises can have a deeper understanding of the feelings and needs of customers.

[0026] Specifically, the question-reply semantic coding unit 323 is used to perform content segmentation and semantic coding on the usage feedback text content to obtain a set of question-reply semantic features. That is, in an embodiment of the present application, first, the usage feedback text content is segmented in units of questions to obtain a set of question-reply pairs; considering that each question in the usage feedback text content contains different emotional label information of the customer, therefore, in order to more accurately understand the customer's specific views and feelings on products and services, in the technical solution of the present application, the usage feedback text content is segmented in units of questions to obtain a set of question-reply pairs. In this way, the customer's feedback can be subdivided into specific question-answer pairs, and the customer feedback of each question can be analyzed more carefully, improving the depth and breadth of the analysis, so that the subsequent emotional analysis of the reply to each question can be performed in a more fine-grained manner, so as to fully understand the customer's feedback in various aspects and form a more complete customer portrait. Then, each question-reply pair in the set of question-reply pairs is semantically coded to obtain a set of question-reply semantic coding feature vectors as the set of question-reply semantic features. Here, considering that each question-answer pair contains the semantic information of the key point question and the answer, which implies the customer's implicit emotional information and content about the product, therefore, in the technical solution of the present application, each question-answer pair in the set of question-answer pairs is semantically encoded to capture and refine the deep semantic information in each text, and obtain a set of question-answer semantic encoding feature vectors, thereby forming a more detailed user portrait, helping enterprises to understand the customer's emotional tendencies on each question in more detail, and improving the accuracy of sentiment analysis.

[0027] Specifically, the satisfaction analysis unit 324 is used to perform a sequence conditional description on the set of question-reply semantic features to distinguish and strengthen to obtain a set of question-reply reinforcement semantic features, and determine whether the target customer is satisfied with the product based on the set of question-reply reinforcement semantic features. Considering that each original question-reply semantic encoding feature vector in the set of question-reply semantic encoding feature vectors may lack sufficient discrimination in some cases, it is difficult for the model to accurately distinguish different questions and replies. For example, different questions and replies may be very similar in vocabulary and grammar, but the actual meaning and emotional tendency are very different. The feature vector lacking discrimination will cause the model to perform poorly in subsequent tasks such as classification and sentiment analysis, so that it is impossible to effectively capture subtle key semantic differences, thereby affecting the accuracy and reliability of the analysis. Therefore, in the technical solution of the present application, the set of question-reply semantic features is described in a sequence conditional manner to distinguish and strengthen to obtain a set of question-reply reinforcement semantic features, so that the feature vector can be dynamically reconstructed to enhance the discrimination, so that it can not only effectively mine the deep semantic information structure from the original data, but also significantly improve the quality of feature representation. This enables the model to more accurately capture the subtle differences and complex relationships between questions and responses, thereby providing more precise analysis results and decision support.

[0028] Specifically, the specific steps of distinguishable reinforcement of the sequence condition description include: first, performing distinguishable reinforcement of the sequence condition description on the set of question-reply semantic encoding feature vectors to obtain a set of question-reply reinforcement semantic encoding feature vectors as the set of question-reply reinforcement semantic features. In a specific example of the present application, first, a field mapper is used to perform field mapping on each question-reply semantic encoding feature vector in the set of question-reply semantic encoding feature vectors. A field mapping operation is performed on each vector in the set of original question-reply semantic encoding feature vectors to convert these features into a form that is more suitable for the current model or task requirements, so as to obtain a set of entry question-reply semantic encoding feature vectors.

[0029] More specifically, a field mapper is used to perform field mapping on each question-response semantic coding feature vector in the set of question-response semantic coding feature vectors using the following field mapping formula to obtain a set of entry question-response semantic coding feature vectors; wherein the field mapping formula is: ;in, The first one in the set of question-answer semantic encoding feature vectors question-response semantic encoding feature vector, and are the first weight matrix and the second weight matrix respectively, yes The corresponding entry question-response semantic encoding feature vector Next, the essential features of the set of entry question-reply semantic coding feature vectors are calculated to obtain the question-reply semantic field essential feature vector. In a specific example of the present application, the field depth factor of each entry question-reply semantic coding feature vector in the set of entry question-reply semantic coding feature vectors is calculated to obtain a set of question-reply semantic field depth factors; based on the set of question-reply semantic field depth factors, the question-reply semantic field essential feature vector of the set of entry question-reply semantic coding feature vectors is calculated. That is, using the set of question-reply semantic field depth factors obtained above as a basis, the field effect normalization mechanism is used to extract a feature distribution that comprehensively reflects the core attributes of the entire batch of entry feature vectors, namely, the question-reply semantic field essential feature vector.

[0030] More specifically, the essential features of the set of the entry question-answer semantic encoding feature vectors are calculated using the following essential feature extraction formula to obtain the question-answer semantic field essential feature vector; wherein the essential feature extraction formula is: ;in, To calculate the square of the vector-norm, represents the logarithmic function value with base 2, For calculation The depth of field factor, for The corresponding question-response semantic field depth factor, It means to calculate the exponential function value with the natural constant e as the base. for The corresponding standardized question-response semantic field depth factor, It is the essential feature vector of the question-response semantic field.

[0031] Furthermore, based on the question-reply semantic field essential feature vector, the set of question-reply semantic coding feature vectors is semantically enhanced to obtain the set of question-reply enhanced semantic coding feature vectors. In a specific example of the present application, the question-reply semantic field essential feature vector is used as a conditional feature vector, and each question-reply semantic coding feature vector in the set of question-reply semantic coding feature vectors is distinguishably enhanced to obtain the set of question-reply enhanced semantic coding feature vectors. That is, using the question-reply semantic field essential feature vector as additional conditional information, the set of question-reply semantic coding feature vectors is subjected to element-by-element distinguishable enhancement processing, so as to specifically amplify important features of certain aspects according to specific rules or weight adjustment schemes, while suppressing other less important parts, so as to achieve the purpose of optimizing feature expression.

[0032] More specifically, based on the question-answer semantic field essential feature vector, the set of question-answer semantic encoding feature vectors is semantically enhanced using the following semantic enhancement formula to obtain the set of question-answer enhanced semantic encoding feature vectors; wherein the semantic enhancement formula is: ;in, is the question-answer semantic essence weight matrix, is the local question-response semantic weight matrix, is the bias vector, yes function, It is the point multiplication by position. Yes Question-answer enhanced semantic encoding feature vector after semantic enhancement.

[0033] Then, after obtaining the set of question-reply reinforcement semantic features, further based on the set of question-reply reinforcement semantic features, a set of question-granularity sentiment polarity labels is obtained. In a specific example of the present application, the set of question-reply reinforcement semantic coding feature vectors is input into a classifier-based question-granularity sentiment polarity identifier to obtain the set of question-granularity sentiment polarity labels. That is, the set of question-reply reinforcement semantic features after reinforcement using the set of question-reply semantic coding feature vectors is classified and processed, so as to intelligently determine the sentiment polarity labels of each question level and determine whether the target customers are satisfied with the product. In this way, a fine-grained analysis of customer usage feedback can be performed to capture the customer's complex emotions, understand the customer's actual feelings and specific problems, and thus accurately judge the customer's sentiment polarity, and achieve a comprehensive and accurate analysis of customer feedback. Specifically, the specific process of inputting the set of question-reply reinforcement semantic coding feature vectors into the classifier-based question-granularity sentiment polarity identifier includes: using multiple fully connected layers of the classifier-based question-granularity sentiment polarity identifier to fully connect the set of question-reply reinforcement semantic coding feature vectors to obtain a set of encoded classification feature vectors; and passing the set of encoded classification feature vectors through the Softmax classification function of the classifier to obtain the set of question-granularity sentiment polarity labels.

[0034] Preferably, inputting the set of question-answer reinforced semantic encoding feature vectors into a classifier-based question-granularity sentiment polarity identifier to obtain a set of question-granularity sentiment polarity labels comprises: All feature values ​​of the set of question-answer reinforcement semantic coding feature vectors are subjected to feature clustering to obtain a feature set within the question-answer reinforcement semantic coding class and a feature set outside the question-answer reinforcement semantic coding class, namely: ;in, represents the feature set within the question-answer reinforced semantic encoding class, Representing the set of out-of-class features of the question-answer reinforcement semantic encoding; The ratio of the number of feature values ​​in the feature set of the question-answer enhanced semantic coding class to the number of all feature values ​​in the set of the question-answer enhanced semantic coding feature vector is calculated, that is: ;in, represents the number of feature values ​​in the feature set within the question-answer reinforcement semantic encoding class, The number of all feature values ​​of the set of feature vectors representing the question-answer reinforcement semantic encoding; Calculate the absolute value of all feature values ​​in the feature set of the question-response reinforcement semantic encoding class The power of the sum of the absolute values ​​of all feature values ​​in the set of question-answer reinforced semantic encoding feature vectors The ratio of the powers is used to obtain the first question-answer reinforcement semantic encoding weight, that is: ;in, Represents each feature value in the set of question-answer reinforced semantic encoding feature vectors, represents the first question-answer reinforcement semantic encoding weight; Calculate the sum of squares of all feature values ​​in the feature set of the question-response reinforcement semantic encoding class The power of the sum of the squares of all eigenvalues ​​in the set of the question-answer reinforced semantic encoding feature vector The ratio of the powers is used to obtain the second question-answer reinforcement semantic encoding weight, namely: ;in, represents the second question-answer reinforcement semantic encoding weight; For each feature value in the feature set within the question-answer enhanced semantic coding class, the product of the feature value and the second question-answer enhanced semantic coding weight is calculated, and then the product of the feature value and the first question-answer enhanced semantic coding weight is added to obtain the optimized feature value, that is: ;in, represents the optimized eigenvalue; For each feature value in the feature set outside the question-answer enhanced semantic coding class, the product of the feature value and the weight of the first question-answer enhanced semantic coding is calculated to obtain the optimized feature value, that is: ;in, represents the optimized eigenvalue; A set of optimized question-answer reinforced semantic coding feature vectors based on the optimized feature values ​​of the question-answer reinforced semantic coding in-class feature set and the question-answer reinforced semantic coding out-class feature set is input into a classifier-based question-granularity sentiment polarity identifier to obtain a set of question-granularity sentiment polarity labels.

[0035] Here, in the case where each question-reply semantic coding feature vector in the set of question-reply semantic coding feature vectors represents the semantic coding feature of the question-reply, when performing feature distinguishable enhancement based on the sequence condition description, considering the sequence condition descriptive differences based on the semantic coding differences, it is expected to improve the semantically consistent aggregate expression effect of the set of question-reply enhanced semantic coding feature vectors obtained by distinguishable enhancement.

[0036] Therefore, while performing feature clustering on the set of question-reply reinforced semantic coding feature vectors, the interaction description of the key feature information of the set of question-reply reinforced semantic coding feature vectors in the clustering process is carried out, and the geometric equivariant topology of the features is constructed by feature low-rank harmonic modulation based on the equivariance of the clustering features of the set of question-reply reinforced semantic coding feature vectors and the features as a whole, so as to obtain the translation and rotation symmetry of the schematic distribution of the clustering features of the set of question-reply reinforced semantic coding feature vectors relative to the features as a whole, thereby introducing geometric message passing in the feature expression of the set of question-reply reinforced semantic coding feature vectors and realizing the cluster mapping symmetry of the set of question-reply reinforced semantic coding feature vectors through manipulation of irreducible low-rank order coefficients, thereby improving the consistency of the clustering-based feature representation of the set of question-reply reinforced semantic coding feature vectors, thereby improving the accuracy of the set of question-granularity sentiment polarity labels obtained by inputting the set of question-reply reinforced semantic coding feature vectors into the classifier-based question-granularity sentiment polarity identifier.

[0037] Finally, based on the set of the problem granularity sentiment polarity labels, it is determined whether the target customer is satisfied with the product. In a specific example of the present application, the number of positive sentiment polarity labels and the number of negative sentiment polarity labels in the set of the problem granularity sentiment polarity labels are counted; and based on the comparison between the number of the positive sentiment polarity labels and the number of the negative sentiment polarity labels, it is determined whether the target customer is satisfied with the product. In particular, customers may express both positive and negative emotions in feedback. For example, a customer may be very satisfied with a certain function of a product, but dissatisfied with another function. By counting the number of positive and negative sentiment polarity labels, these emotions can be balanced, avoiding the bias of a single dimension, thereby obtaining a more objective satisfaction evaluation result. Based on the statistical results of the sentiment polarity labels, enterprises can make decisions more scientifically. In one example, in response to the number of the positive sentiment polarity labels being greater than the number of the negative sentiment polarity labels, the feedback result is that the target customer is satisfied with the product; otherwise, the feedback result is that the target customer is dissatisfied with the product, and the enterprise can make improvements to the product to improve customer satisfaction.

[0038] In particular, the enterprise management module 330 is used to monitor the enterprise's market share and competitive product dynamics, and evaluate employees' sense of identity with the enterprise. In the enterprise intelligent management system, monitoring the enterprise's market share and competitive product dynamics, as well as evaluating employees' sense of identity with the enterprise, are key links to ensure the healthy and sustainable development of the enterprise. In one example, first of all, data collection is the basis of the entire monitoring process. Enterprises need to collect market data through various channels, including industry reports, data from market research institutions, government statistics, etc. These data can help enterprises understand the size, growth rate and trend of the entire market. In addition, enterprises also need to collect relevant information about competitors, including product features, pricing strategies, marketing activities, customer reviews, etc. This information can be obtained through web crawlers, social media monitoring, industry forums and professional media. For example, you can use a web crawler tool to crawl the official website information of competitors, monitor users' evaluation and discussion of competitors' products through social media monitoring tools, subscribe to industry news and forums, and obtain the latest competitor dynamics. At the same time, enterprises should also collect customer feedback on the company and competitors' products through customer satisfaction surveys, online comments and social media monitoring, and understand customers' views and preferences on different products. Next, data collation and analysis are key steps to ensure data quality and accuracy of analysis results. Enterprises need to clean and standardize the collected data, remove invalid or duplicate data, and ensure the accuracy and consistency of the data. Data cleaning scripts can be written in programming languages ​​such as Python to automate data cleaning tasks. Based on market data and enterprise sales data, enterprises can calculate their share of the target market. Market share can be measured by indicators such as sales volume and sales volume.

[0039] In addition, companies also need to compare and analyze their products with those of their main competitors, including price, performance, quality, service and other aspects. The advantages and disadvantages of corporate products and competing products can be quantified by building a comparison table or scoring model. For example, a scoring table can be designed to score the products of companies and competitors in terms of price, performance, quality and service. The full score for each aspect can be set to 10 points, and then the total score is calculated for comparative analysis. Companies also need to analyze market data to identify market trends and changes, including emerging markets, changes in consumer demand, and technological innovation. This helps companies adjust their market strategies in a timely manner and seize market opportunities. Excel or Python can be used for data analysis, trend charts, and identification of long-term trends and short-term fluctuations in the market.

[0040] Finally, data visualization and report generation are important steps to present the analysis results to management. Enterprises can display the analysis results in the form of charts, dashboards, etc., so that managers can intuitively understand the market share and competitive dynamics. Common visualization tools include Tableau, Power BI, etc. For example, Tableau can be used to create a dashboard of market share and competitive dynamics to show the market performance of enterprises and competitors, including market share, sales, sales volume and other indicators. Enterprises should also regularly generate market analysis reports to summarize the market status, competitor dynamics and enterprise market performance, and put forward improvement suggestions and market strategies. The report should include data sources, analysis methods and conclusions to ensure the transparency and credibility of information. For example, a market analysis report can be generated every month to summarize the market performance of the previous month, analyze the dynamics of competitors, and put forward the next market strategy and improvement suggestions.

[0041] Employee data collection is the basis for evaluating employee identity. Companies need to collect basic information about employees, including name, position, time of joining, department, etc., to provide basic data for subsequent analysis. Basic information about employees can be exported from the company's human resources management system (such as SAP HR) to ensure the accuracy and completeness of the data. Companies also need to collect employees' opinions on their recognition of corporate culture, satisfaction with the work environment, and satisfaction with compensation and benefits through regular employee satisfaction surveys. Surveys can be conducted through online questionnaires, paper questionnaires, or face-to-face interviews. For example, online questionnaire platforms (such as SurveyMonkey) can be used to conduct regular employee satisfaction surveys, design questionnaire questions, including multiple-choice questions and open-ended questions, and ensure that the questionnaire content fully covers all aspects of employee concern. At the same time, companies also need to collect employee work performance data, including work completion rate, project results, customer evaluation, etc., to evaluate employee work performance and contribution. Employee performance data can be exported from the company's performance management system (such as Workday) to ensure the accuracy and timeliness of the data.

[0042] Next, data collation and analysis are key steps to ensure data quality and the accuracy of analysis results. Companies need to clean and standardize the collected employee data, remove invalid or duplicate data, and ensure data accuracy and consistency. Data cleaning scripts can be written in programming languages ​​such as Python to automate data cleaning tasks. Based on employee length of service and work performance data, companies can calculate employee loyalty.

[0043] Companies also need to analyze the results of employee satisfaction surveys and collect employee satisfaction scores in terms of corporate culture, work environment, compensation and benefits. The overall employee satisfaction can be evaluated by calculating statistical indicators such as average score and standard deviation. For example, Excel or Python can be used to analyze the results of employee satisfaction surveys, draw bar charts and pie charts, and show the distribution of employee satisfaction in various aspects.

[0044] Finally, companies also need to integrate employee loyalty and satisfaction data to assess employees' sense of identification with the company.

[0045] Companies can display employee identity assessment results in the form of charts, dashboards, etc., so that managers can intuitively understand the overall status of employees. Commonly used visualization tools include Tableau, Power BI, etc. For example, Tableau can be used to create an employee identity assessment dashboard to display employee satisfaction scores in terms of corporate culture, work environment, compensation and benefits, as well as the overall identity value distribution of employees. Companies should also regularly generate employee identity assessment reports to summarize employees' identification with corporate culture, satisfaction with the work environment, and satisfaction with compensation and benefits, and put forward improvement suggestions and management strategies. The report should include data sources, analysis methods and conclusions to ensure the transparency and credibility of the information. For example, an employee identity assessment report can be generated every quarter to summarize the employee satisfaction and identity level of the previous quarter, analyze existing problems and deficiencies, and put forward specific improvement suggestions and management strategies.

[0046] Through the above steps, companies can effectively monitor market share and competitive product dynamics, evaluate employees' identification with the company, thereby promptly discovering and solving problems, optimizing market strategies and internal management, and ensuring the healthy and sustainable development of the company.

[0047] In particular, the risk analysis module 340 is used to analyze and predict the risks of the enterprise operation based on the customer's satisfaction with the product and the employee's identification with the enterprise, and obtain the analysis results. In enterprise management and operation, customer satisfaction with the product and employee identification with the enterprise are two crucial indicators. Through the analysis and prediction of these two indicators, the enterprise can have a more comprehensive understanding of its own operating conditions, timely discover and solve potential risks, and thus ensure the healthy and sustainable development of the enterprise. It should be understood that customer satisfaction with the product directly affects customer loyalty. Specifically, enterprises with high customer satisfaction often have a better brand image, positive customer feedback and evaluation can enhance the company's visibility and reputation, attract more potential customers, and high customer satisfaction can become a competitive advantage for the enterprise, making the enterprise stand out in the fierce market competition; while enterprises with low customer satisfaction may face negative market evaluation and public opinion pressure, affecting the company's market reputation, and even leading to customer loss and market share decline. Employees' identification with corporate culture and satisfaction with their work directly affect their loyalty. More specifically, loyal employees are more willing to stay in the company for a long time, reducing the costs and risks brought by talent loss; a stable workforce helps to maintain the continuity and stability of the company. Therefore, by analyzing and predicting the risks of business operations based on customer satisfaction with products and employee identification with the company, it can not only help companies to promptly discover and solve potential problems, optimize market strategies and internal management, but also improve customer satisfaction and employee loyalty, and enhance the company's competitiveness and market position. This provides a solid foundation for the healthy and sustainable development of the company.

[0048] In particular, the early warning processing module 350 is used to generate corresponding early warnings and formulate corresponding emergency strategies based on the analysis results. In one example, by monitoring customer satisfaction, the company can promptly discover problems in products or services, take measures to improve them, and avoid customer loss and market share loss due to quality problems; by monitoring employee satisfaction, the company can promptly discover problems in the working environment, salary and benefits, and take measures to improve employees' working conditions and improve their job satisfaction and loyalty. In this way, not only can potential risks be discovered and responded to in a timely manner, operational decisions can be optimized, customer and employee satisfaction can be improved, and corporate competitiveness and market position can be enhanced, but also operational risks can be effectively reduced to ensure the healthy and sustainable development of the company.

[0049] As described above, the cloud-based enterprise intelligent management system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a cloud-based enterprise intelligent management algorithm. In a possible implementation, the cloud-based enterprise intelligent management system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the cloud-based enterprise intelligent management system 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the cloud-based enterprise intelligent management system 300 can also be one of the many hardware modules of the wireless terminal.

[0050] Alternatively, in another example, the cloud platform-based enterprise intelligent management system 300 and the wireless terminal may also be separate devices, and the cloud platform-based enterprise intelligent management system 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0051] Furthermore, a cloud platform-based enterprise intelligent management method is also provided.

[0052] Figure 4 FIG. 1 is a flow chart of a cloud platform-based enterprise intelligent management method according to an embodiment of the present application. Figure 4 As shown, according to the cloud platform-based intelligent enterprise management method of the embodiment of the present application, the steps include: S1, storing the collected customer usage feedback questionnaires; S2, analyzing the customer satisfaction with the product based on the customer usage feedback questionnaires; S3, monitoring the enterprise market share and competitive product dynamics, and evaluating the employees' sense of identification with the enterprise; S4, based on the customer satisfaction with the product and the employees' sense of identification with the enterprise, analyzing and predicting the risks of the enterprise operation to obtain analysis results; S5, based on the analysis results, generating corresponding warnings and formulating corresponding emergency strategies.

[0053] In summary, the cloud platform-based intelligent enterprise management method according to the embodiment of the present application is explained, which extracts the target customer's usage feedback questionnaire from the questionnaire storage module and extracts the usage feedback text content, and then uses natural language processing technology and encoding methods based on deep learning to perform segmentation and semantic encoding of the usage feedback text content, thereby conditionally describing and reinforcing the semantic encoding features of each encoded question-response semantic code to achieve intelligent judgment of the emotional polarity label of each question level, and determine whether the target customer is satisfied with the product. Compared with a simple scoring method, in this way, a fine-grained analysis of customer usage feedback can be performed to capture the customer's complex emotions, understand the customer's actual feelings and specific problems, thereby accurately judging the customer's emotional polarity and achieving a comprehensive and accurate analysis of customer feedback.

[0054] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. An enterprise intelligent management system based on a cloud platform, characterized in that: The enterprise intelligent management system is deployed on a cloud platform. The enterprise intelligent management system based on the cloud platform includes: A questionnaire storage module is used to store the collected customer usage feedback questionnaires; a customer management module is used to analyze customer satisfaction with the product based on the customer usage feedback questionnaires; an enterprise management module is used to monitor the enterprise's market share and competitive product dynamics, and evaluate employees' sense of identity with the enterprise; a risk analysis module is used to analyze and predict the risks of the enterprise's operations based on customer satisfaction with the product and employees' sense of identity with the enterprise, and obtain analysis results; an early warning processing module is used to generate corresponding early warnings and formulate corresponding emergency strategies based on the analysis results; wherein the customer management module includes: A questionnaire extraction unit, used for extracting the target customer's usage feedback questionnaire from the questionnaire storage module; A feedback text content extraction unit, used to extract usage feedback text content from the usage feedback questionnaire of the target customer; A question-answer semantic encoding unit, used for performing content segmentation and semantic encoding on the usage feedback text content to obtain a set of question-answer semantic features; The satisfaction analysis unit is used to perform sequence condition description distinguishable reinforcement on the set of question-answer semantic features to obtain a set of question-answer reinforcement semantic features, and determine whether the target customer is satisfied with the product based on the set of question-answer reinforcement semantic features.

2. The cloud platform-based enterprise intelligent management system according to claim 1, characterized in that: The question-response semantic encoding unit comprises: A content segmentation subunit, used to segment the usage feedback text content based on questions to obtain a set of question-answer pairs; The semantic encoding subunit is used to perform semantic encoding on each question-answer pair in the set of question-answer pairs to obtain a set of question-answer semantic encoding feature vectors as the set of question-answer semantic features.

3. The cloud platform-based enterprise intelligent management system according to claim 2, characterized in that: The satisfaction analysis unit comprises: A question-answer semantic enhancement unit, used for performing sequence condition description distinguishable enhancement on the set of question-answer semantic encoding feature vectors to obtain a set of question-answer enhancement semantic encoding feature vectors as the set of question-answer enhancement semantic features; A sentiment polarity label generation unit, used to obtain a set of question-granularity sentiment polarity labels based on the set of question-answer reinforced semantic features; The satisfaction judgment unit is used to determine whether the target customer is satisfied with the product based on the set of question granularity sentiment polarity labels.

4. The cloud platform-based enterprise intelligent management system according to claim 3, characterized in that: The question-answer semantic reinforcement unit includes: A field mapping subunit, configured to use a field mapper to perform field mapping on each question-response semantic encoding feature vector in the set of question-response semantic encoding feature vectors to obtain a set of entry question-response semantic encoding feature vectors; A question-answer essential feature calculation subunit, used to calculate the essential features of the set of entry question-answer semantic encoding feature vectors to obtain a question-answer semantic field essential feature vector; The essential feature guided semantic reinforcement subunit is used to semantically reinforce the set of question-reply semantic encoding feature vectors based on the essential feature vector of the question-reply semantic field to obtain the set of question-reply reinforced semantic encoding feature vectors.

5. The cloud platform-based enterprise intelligent management system according to claim 4, characterized in that: The question-answer essential feature calculation subunit is used to: Calculating the field depth factor of each entry question-response semantic encoding feature vector in the set of entry question-response semantic encoding feature vectors to obtain a set of question-response semantic field depth factors; Based on the set of question-response semantic field depth factors, the question-response semantic field essential feature vector of the set of entry question-response semantic encoding feature vectors is calculated.

6. The cloud platform-based enterprise intelligent management system according to claim 5, characterized in that: The essential feature guided semantic reinforcement subunit is used to: use the essential feature vector of the question-reply semantic field as a conditional feature vector, and respectively perform distinguishable reinforcement on each question-reply semantic coding feature vector in the set of question-reply semantic coding feature vectors to obtain the set of question-reply reinforcement semantic coding feature vectors.

7. The cloud platform-based enterprise intelligent management system according to claim 6, characterized in that: The sentiment polarity label generation unit is used to: input the set of question-answer reinforced semantic encoding feature vectors into a classifier-based question granularity sentiment polarity identifier to obtain the set of question granularity sentiment polarity labels.

8. The cloud platform-based enterprise intelligent management system according to claim 7, characterized in that: The satisfaction judgment unit is used to: Counting the number of positive sentiment polarity labels and the number of negative sentiment polarity labels in the set of question-granularity sentiment polarity labels; Based on the comparison between the number of the sentiment positive polarity labels and the number of the sentiment negative polarity labels, it is determined whether the target customer is satisfied with the product.

9. A cloud platform-based enterprise intelligent management method, characterized in that: include: Store the collected customer feedback questionnaires; Analyze customer satisfaction with the product based on the customer feedback questionnaire; Monitor the company's market share and competitive dynamics, and assess employees' identification with the company; Based on customer satisfaction with the product and employees' identification with the enterprise, analyze and predict the risks of the enterprise's operations and obtain analysis results; Based on the analysis results, corresponding warnings are generated and corresponding emergency strategies are formulated.

Images