Public opinion big data driven kindergarten management intelligent system

Abstract

The invention relates to the technical field of education and artificial intelligence application, in particular to a public opinion big data driven kindergarten management intelligent system, which comprises a data collection module, a data analysis module, an optimal case library construction module, a question classification module, an automatic reply generation module, a public opinion monitoring module and a satisfaction evaluation module. Wherein the data collection module is used for collecting interactive data between parents and kindergartens, including e-mails, short messages and social media messages, and transmitting the data to the central database for unified management; and the data analysis module is connected with the central database and is used for extracting themes, emotions and information points of the questions so as to evaluate the importance and the emergency degree of the questions. According to the invention, through real-time monitoring of social media and online feedback, automatic reply generation and deep satisfaction evaluation, the response efficiency of the kindergarten and the satisfaction of parents are significantly improved, and the public image of the kindergarten is effectively maintained and improved.

Description

Technical Field

[0001] The present invention relates to the field of education and artificial intelligence application technology, and in particular to a kindergarten management intelligent system driven by public opinion big data. Background Art

[0002] With the advent of the digital age, kindergartens, as a bridge between families and schools, are faced with the challenge of improving parent communication efficiency and satisfaction. Modern parents expect instant and transparent communication methods, which requires kindergartens to be more efficient and accurate in handling parent feedback and monitoring public image. However, existing kindergarten management systems often lack effective tools to analyze large amounts of parent feedback data, especially real-time data from social media and other digital platforms. In addition, traditional communication response mechanisms cannot quickly adapt to parents' specific needs and emergencies, resulting in damage to parent satisfaction and the kindergarten's public image.

[0003] The main problem with existing technologies is the lack of a comprehensive system to monitor and analyze parent feedback from different channels in real time, and to provide fast and personalized responses based on this. Existing solutions generally fail to make full use of advanced data analysis and natural language processing technologies, and are therefore unable to accurately identify and prioritize key information, such as urgent feedback or issues that have a significant impact on kindergarten policies.

[0004] Therefore, there is an urgent need for a kindergarten management system that can integrate the latest information technology to improve the efficiency of handling parent interactions, optimize the response process, and maintain and enhance the public image of the kindergarten through accurate public opinion monitoring. Summary of the invention

[0005] Based on the above objectives, the present invention provides an intelligent kindergarten management system driven by public opinion big data.

[0006] The kindergarten management intelligent system driven by public opinion big data includes data collection module, data analysis module, best case library construction module, question classification module, automatic reply generation module, public opinion monitoring module and satisfaction evaluation module; among which:

[0007] Data collection module: used to collect interaction data between parents and kindergartens, including emails, text messages and social media messages, and transmit the data to the central database for unified management;

[0008] Data analysis module: connects to the central database and applies preset natural language processing technology to analyze the collected data, extracting the topic, sentiment and information points of the problem to evaluate the importance and urgency of the problem;

[0009] Best Case Library Building Module: Based on the output of the data analysis module, build and update the case library containing standard responses and best practice examples;

[0010] Question Classification Module: Classify and prioritize questions based on the case library provided by the Best Case Library Building Module, and generate standardized response suggestions;

[0011] Automatic reply generation module: Generates customized reply content based on the standardized reply suggestions provided by the question classification module to support effective communication between teachers and parents;

[0012] Public opinion monitoring module: real-time monitoring and analysis of public feedback from social media and online platforms to assess the kindergarten’s public image and public opinion dynamics;

[0013] Satisfaction evaluation module: Based on parents’ feedback and interaction results, evaluate parent satisfaction and provide data support for improving educational services.

[0014] Furthermore, the data collection module includes a receiving unit, a preprocessing unit, a classification unit and a transmission unit; wherein:

[0015] Receiving unit: adopts a multi-protocol interface design to be compatible with different communication platforms and formats, and is used to receive interactive data between parents and kindergartens through email, SMS and social media channels. The multi-protocol interface includes SMTP protocol for email reception, SMS gateway for SMS transmission, and social media APIs for capturing social media exchanges. The multi-protocol interface can automatically identify and receive data from different electronic communication tools and cache them;

[0016] Preprocessing unit: Perform data format standardization and cleaning operations through preset automated tools, including removing invalid and irrelevant information, correcting format errors, and unifying data coding;

[0017] Classification unit: Use preset rules to preliminarily classify the preprocessed data for subsequent processing. The specific classification is based on the basic attributes of the data type and content. The classification automatically labels each piece of data through the classification algorithm;

[0018] Transmission unit: Encryption technology and data transmission protocol are used to transmit classified and pre-processed data to the central database. The specific technologies used include TLS / SSL encryption protocol to ensure the security and integrity of data during transmission and prevent data from being illegally accessed or tampered with during transmission.

[0019] Furthermore, the data analysis module includes a connection unit, a natural language processing unit and a preliminary evaluation unit; wherein:

[0020] Connection unit: Uses standard database connection protocols to establish a stable connection with the central database, providing instant data flow for natural language processing and subsequent analysis;

[0021] Natural language processing unit: configures preset natural language processing technology to parse and understand text data transmitted from the data collection module. Specific technologies include semantic analysis, sentiment analysis, and core word extraction to extract the topic, sentiment, and information points of the question from the text;

[0022] Preliminary evaluation unit: Use decision tree technology to perform preliminary importance and urgency evaluation on the data obtained from the natural language processing unit to screen out the data to be processed in priority and mark the screened data as high priority.

[0023] Furthermore, the natural language processing unit includes:

[0024] Text preprocessing: First, the received text data is cleaned and standardized, including removing useless characters, punctuation marks, and converting all characters to lowercase, in preparation for more in-depth language processing, ensuring the consistency and cleanliness of the text data, and laying the foundation for subsequent analysis;

[0025] Semantic analysis technology: The bag-of-words model and TF-IDF weight calculation are used to identify core words and phrases in the text. The calculation formula of TF-IDF is: TF-IDF(t,d)=TF(t,d)×IDF(t), where TF(t,d) is the frequency of occurrence of term t in document d, and IDF(t) is the inverse document frequency. The calculation method is: is the total number of documents, and the denominator is the number of documents containing the term;

[0026] Sentiment analysis: Use VADER's sentiment analysis algorithm to assess the sentiment tendency of the text. VADER is a sentiment analysis tool based on dictionaries and rules, which is used for social media texts. It associates each word in the text with a sentiment score and calculates the sentiment score of the entire text. The sentiment score formula is: Sentiment score = ∑ (term score) × modifier, where the modifier is adjusted according to the negative words and emphasis words before the word to enhance the accuracy of the analysis;

[0027] Information point extraction: Combine natural language processing technology and named entity recognition methods to identify and extract specific information points from the text, including names, places, and dates.

[0028] Further, the preliminary assessment unit includes:

[0029] Decision tree construction: First, a decision tree model is built based on historical data and expert knowledge. The model aims to evaluate the importance and urgency of the question based on the characteristics of the input data, the question topic, sentiment score, and information points obtained from the natural language processing unit;

[0030] Feature extraction: Extracting features required by the decision tree model from the data received by the natural language processing unit. Specific features include topic relevance score, sentiment tendency, and urgency index;

[0031] Decision tree evaluation algorithm: Apply the decision tree algorithm to evaluate the extracted features and output the importance and urgency rating of each problem. The evaluation process formula of the decision tree is: Y = f(X), where X is the vector of input features and Y is the output rating, indicating the priority of problem handling;

[0032] Priority marking step: Based on the ratings output by the decision tree, the data is automatically marked into different priority categories, and high priority issues will be marked as urgent.

[0033] Furthermore, the best case library construction module includes a case analysis unit, a case construction unit, a case update unit and a case storage unit; wherein:

[0034] Case analysis unit: used to analyze the data received from the data analysis module, including the analysis results of the topic, sentiment and urgency of the questions, and identify the frequently occurring questions with high importance or urgency based on the analysis results;

[0035] Case Building Unit: Based on the analysis results of the case analysis unit, new cases are constructed, including the development of standard responses and the definition of best practice examples. Each example is used to address frequently occurring parent issues. Each new case will detail the background of the issue, the recommended response measures, the expected results, and the follow-up actions.

[0036] Case Update Unit: used to monitor the validity and timeliness of cases in the existing case library, regularly review the case library, and adjust or update the case content based on the latest data analysis results to ensure that the case library always reflects current parent issues and effective response strategies;

[0037] Case Storage Unit: Saves constructed or updated cases in a structured database for quick retrieval and use.

[0038] Furthermore, the question classification module includes a classification decision unit, a priority assessment unit and a reply suggestion generation unit; wherein:

[0039] Classification Decision Unit: Uses a rule-based algorithm to classify questions. Each new question is classified into predefined categories based on data extracted from the case library provided by the Best Case Library Building Module, including the type of question, urgency indicator, and common problem patterns. The rule-based algorithm classifies questions based on their similarity to cases in the case library. The calculation formula specifically uses cosine similarity to determine the degree of match between questions and known categories: Similarity in, is the eigenvector of the new problem, is the feature vector of the category in the case library;

[0040] Priority Assessment Unit: Assess the priority of the problem based on its urgency and impact. Specifically, a weight allocation method is used to assign a priority score to each problem based on the urgency index, emotional intensity, and scope of the problem. The specific calculation formula is:

[0041] Priority score (p) = w 1 ×I+w 2 ×E+w 3 ×S, where I is the impact scope of the problem, E is the emotional intensity, S is the urgency of the problem, and w 1 ,w 2 ,w 3 is the weight of each factor;

[0042] Response suggestion generation unit: Based on the classification results and priority assessment, standardized response suggestions are automatically generated. The response suggestions are guided by the best practices of similar questions in the case library to ensure that the answers provided are in line with the kindergarten's policies and meet the specific needs of parents.

[0043] Furthermore, the automatic reply generation module includes a reply template selection unit, a content customization unit, a language optimization unit and a reply review unit; wherein:

[0044] Response template selection unit: The Jaccard similarity algorithm is used to select the template that best matches the current question from the standardized response suggestion library. The Jaccard similarity evaluates the similarity by comparing the ratio of the intersection and union of the question feature set and the template feature set. The specific calculation formula is: Among them, A represents the feature set of the question, B represents the feature set of the template, and the Jaccard similarity value ranges from 0 to 1. The higher the value, the higher the similarity between the two sets, so the most similar reply template is selected;

[0045] Content customization unit: After selecting the basic response template, customize the response according to the details and context of the specific question, making the response more personal and specific by adding a scheduled date, event name or targeted suggestions;

[0046] Language Optimization Unit: used to adjust the language and style of responses to ensure that they comply with the kindergarten's communication guidelines and are easy for parents to understand. Language processing technology is used to optimize the tone and politeness.

[0047] Response Review Unit: Before the response is sent, the generated response content is subject to final review. During the review process, the accuracy, completeness and sensitivity of the response are checked to ensure that all responses can accurately convey the required information.

[0048] Furthermore, the public opinion monitoring module includes a data collection unit, a real-time analysis unit, a trend analysis unit and a report generation unit; wherein:

[0049] Data collection unit: used to continuously monitor and collect public feedback from social media and other online platforms, specifically using APIs and web crawler technology to automatically capture posts and comments mentioning kindergartens from platforms and online forums;

[0050] Real-time analysis unit: Apply stream data processing technology, including Apache Kafka or Apache Storm, to analyze the collected data. The real-time analysis tasks include sentiment analysis and core word recognition of the text. The sentiment score is calculated using the following formula: Among them, sentiment value i is the numerical value of the sentiment tag extracted from a single post, and weight i is the weight determined based on word frequency and contextual importance;

[0051] Trend analysis unit: Use time series analysis technology to analyze the changing trends of public opinion dynamics. By calculating the average and standard deviation of sentiment scores and mention frequencies within a predetermined time window, it is used to identify changing trends and abnormal fluctuations in public opinion. The specific formula is:

[0052] Where T is the time window length, and sentiment score t is the sentiment score at time t;

[0053] Report generation unit: Based on the analysis results, a comprehensive report on the kindergarten’s public image and public opinion status is generated. The report includes a graphical display of indicators, a summary of focus points and suggested measures to support the kindergarten management in making information-based decisions.

[0054] Furthermore, the satisfaction evaluation module includes a data collection unit, a satisfaction analysis unit, a trend monitoring unit and a report generation unit; wherein:

[0055] Data collection unit: used to collect parents’ feedback and interaction data with the kindergarten, including parents’ comments, ratings and direct feedback on different platforms;

[0056] Satisfaction Analysis Unit: Statistical analysis techniques and machine learning algorithms are used to evaluate parents’ satisfaction. Specifically, the following calculation formula is used to analyze the sentiment tendency and satisfaction score of parents’ feedback:

[0057] Satisfaction score = α × average sentiment score + β × positive feedback ratio + γ × engagement index, where α, β, and γ are adjustment parameters optimized based on historical data to balance the impact of various factors; the average sentiment score is based on sentiment analysis results, the positive feedback ratio is the ratio of the number of positive feedback to the total number of feedback, and the engagement index is calculated based on the frequency and diversity of parent interactions;

[0058] Trend monitoring unit: Use time series analysis to monitor the long-term trend and seasonal changes of satisfaction. Specifically, use the preset autoregressive moving average model to predict future satisfaction trends and provide forward-looking data support for kindergarten management. The specific formula is:

[0059] in, and θ are model parameters, ∈ t is the error term, representing random fluctuations;

[0060] Report generation unit: Generates detailed satisfaction reports based on the analysis results, including indicator charts, trend analysis results and improvement suggestions. The satisfaction report provides decision support for kindergarten management.

[0061] Beneficial effects of the present invention:

[0062] The present invention, through integrated natural language processing and real-time data analysis technology, the system can quickly identify key public opinion trends and potential problems, so that the kindergarten can respond in time and effectively prevent or mitigate possible negative impacts. This advanced monitoring function ensures that the kindergarten can maintain a good image in public opinion while enhancing parents' trust and satisfaction.

[0063] The present invention provides fast and customized response capabilities through an automatic reply generation module, which greatly improves the communication efficiency between kindergartens and parents. The reply module intelligently generates targeted reply suggestions based on data obtained from question classification and best case libraries, ensuring that every communication is accurate and timely. This not only reduces the workload of faculty and staff, but also optimizes the interactive experience of parents, allowing them to feel the attention and professionalism of the kindergarten, thereby enhancing parents' sense of participation and satisfaction.

[0064] The present invention, through the satisfaction evaluation module, can comprehensively analyze parents' feedback and interaction data, accurately measure and track parents' satisfaction. This function not only helps kindergartens understand parents' specific needs and expectations, but also provides data-based insights to guide kindergartens to make effective improvements in education and services. In addition, the detailed satisfaction reports regularly generated by the system provide decision-making support for kindergarten management, enabling them to formulate or adjust strategies based on real-time data and continuously improve the quality and effectiveness of educational services. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 A schematic diagram of an intelligent kindergarten management system according to an embodiment of the present invention;

[0067] Figure 2 Schematic diagram of a public opinion monitoring module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0068] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0069] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. The words "include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0070] like Figure 1-Figure 2 As shown in the figure, the kindergarten management intelligent system driven by public opinion big data includes data collection module, data analysis module, best case library construction module, question classification module, automatic reply generation module, public opinion monitoring module and satisfaction evaluation module; among which:

[0071] Data collection module: used to collect interaction data between parents and kindergartens, including emails, text messages and social media messages, and transmit the data to the central database for unified management;

[0072] Data analysis module: connects to the central database and applies preset natural language processing technology to analyze the collected data, extracting the topic, sentiment and information points of the problem to evaluate the importance and urgency of the problem;

[0073] Best case library building module: Based on the output of the data analysis module, build and update the case library containing standard responses and best practice examples for use by the question classification module;

[0074] Question Classification Module: Classify and prioritize questions based on the case library provided by the Best Case Library Building Module, and generate standardized response suggestions;

[0075] Automatic reply generation module: Generates customized reply content based on the standardized reply suggestions provided by the question classification module to support effective communication between teachers and parents;

[0076] Public opinion monitoring module: real-time monitoring and analysis of public feedback from social media and online platforms to assess the kindergarten’s public image and public opinion dynamics;

[0077] Satisfaction evaluation module: Based on parents’ feedback and interaction results, evaluate parent satisfaction and provide data support for improving educational services.

[0078] The data collection module includes a receiving unit, a preprocessing unit, a classification unit and a transmission unit; wherein:

[0079] Receiving unit: A multi-protocol interface is designed to be compatible with different communication platforms and formats. It is used to receive interactive data between parents and kindergartens via email, text messages and social media channels. The multi-protocol interface includes SMTP protocol for email reception, SMS gateway for text message transmission, and social media APIs (such as Facebook, Twitter, etc.) for capturing social media exchanges. The multi-protocol interface can automatically identify and receive data from different electronic communication tools and cache them;

[0080] Pre-processing unit: Perform data format standardization and cleaning operations through preset automated tools, including removing invalid and irrelevant information, correcting format errors, and unifying data coding to ensure that the data has a consistent format and high quality before entering the analysis phase;

[0081] Classification unit: Use preset rules to perform preliminary classification of pre-processed data for subsequent processing. The specific classification is based on the data type (such as email, SMS, social media message) and the basic attributes of the content (such as the type of question such as safety and health consultation). The classification automatically labels each data through the classification algorithm to simplify the data structure for in-depth analysis and specific applications;

[0082] Transmission unit: Encryption technology and data transmission protocol are used to transmit the classified and pre-processed data to the central database. The specific technologies used include TLS / SSL encryption protocol to ensure the security and integrity of data during transmission and prevent data from being illegally accessed or tampered with during transmission. Through the detailed description of the above data collection module, the system can effectively manage and optimize the communication data between parents and kindergartens. The multi-protocol interface design of the receiving unit ensures the comprehensive reception and efficient caching of various communication data. The preprocessing unit improves the data quality through standardization and cleaning, providing an accurate basis for subsequent data analysis. The setting of the preliminary classification unit makes the data processing process more efficient and easy to manage. Finally, the encryption measures of the transmission unit ensure the security of data transmission, enhance the data protection capabilities of the entire system, and improve the safety and reliability of operations, thereby greatly improving the management efficiency of the kindergarten and the satisfaction of parents.

[0083] The data analysis module includes a connection unit, a natural language processing unit, and a preliminary evaluation unit; wherein:

[0084] Connection unit: Use standard database connection protocols to establish a stable connection with the central database. Through this connection, the data analysis module can access the data stored in the database in real time, providing instant data flow for natural language processing and subsequent analysis;

[0085] Natural language processing unit: configures preset natural language processing technology to parse and understand text data transmitted from the data collection module. Specific technologies include semantic analysis, sentiment analysis, and core word extraction to extract the topic, sentiment, and information points of the question from the text;

[0086] Preliminary evaluation unit: Use decision tree technology to conduct a preliminary importance and urgency assessment of the data obtained from the natural language processing unit to screen out the data to be processed first, and mark the screened data as high priority, ready to be sent to the problem classification module for more detailed processing; through the detailed configuration of the data analysis module, this system can accurately and effectively process the interaction data between the kindergarten and parents. The connection unit ensures the stability and security of the data flow, providing a solid foundation for data analysis. The NLP unit, through its in-depth text analysis capabilities, not only captures the literal meaning of the information, but also analyzes the emotions and topic importance behind it, providing necessary input for decision-making. Finally, the decision tree method of the evaluation unit makes the handling of problems more scientific and systematic, improves the efficiency and accuracy of problem responses, and enhances the kindergarten's ability to prevent and manage potential problems.

[0087] The natural language processing unit includes:

[0088] Text preprocessing: First, the received text data is cleaned and standardized, including removing useless characters, punctuation marks, and converting all characters to lowercase, in preparation for more in-depth language processing, ensuring the consistency and cleanliness of the text data, and laying the foundation for subsequent analysis;

[0089] Semantic analysis technology: Apply the Bag of Words (BoW) model and TF-IDF (Term Frequency-Inverse Document Frequency) weight calculation to identify the core words and phrases in the text. The calculation formula of TF-IDF is: TF-IDF(t, d) = TF(t, d) × IDF(t), where TF(t, d) is the frequency of occurrence of term t in document d, and IDF(t) is the inverse document frequency. The calculation method is: is the total number of documents, and the denominator is the number of documents containing the term. This method helps highlight terms that are important in a specific text but uncommon in other documents, effectively extracting question topics;

[0090] Sentiment analysis: Use the sentiment analysis algorithm VADER (Valence Aware Dictionary ands Entiment Reasoner) to assess the sentiment tendency of the text. VADER is a sentiment analysis tool based on dictionaries and rules, which is used for social media texts. It associates each word in the text with a sentiment score and calculates the sentiment score of the entire text. The sentiment score formula is: Sentiment score = ∑ (term score) × modifier, where the modifier is adjusted according to the negative words and emphasis words before the word to enhance the accuracy of the analysis;

[0091] Information point extraction: Combining natural language processing technology and named entity recognition (NER) methods, specific information points are identified and extracted from the text, including names, places, and dates. NER is usually implemented using machine learning models such as conditional random fields (CRFs), which are trained to recognize named entities in text.

[0092] The initial assessment units include:

[0093] Decision tree construction: First, a decision tree model is built based on historical data and expert knowledge. The model aims to evaluate the importance and urgency of the question based on the characteristics of the input data, the question topic, sentiment score, and information points obtained from the natural language processing unit;

[0094] Feature extraction: Extracting features required by the decision tree model from the data received by the natural language processing unit. The specific features include topic relevance score, sentiment tendency, and urgency index; among them:

[0095] Thematic relevance score: measures the relevance of the text topic to current important issues in kindergarten;

[0096] Sentiment: Analyze the sentiment polarity and intensity of the text to determine whether it is positive, negative or neutral;

[0097] Urgency index: Assess the urgency of information based on time-sensitive words (such as "immediately" and "urgent") in the text.

[0098] Decision tree evaluation algorithm: Apply the decision tree algorithm to evaluate the extracted features and output the importance and urgency rating of each problem. The evaluation process formula of the decision tree is: Y = f(X), where X is the vector of input features and Y is the output rating, indicating the priority of problem handling;

[0099] Priority marking step: According to the rating output by the decision tree, the data is automatically marked into different priority categories. High priority issues will be marked as urgent, and the urgent mark will trigger the system's rapid response mechanism. Through the detailed definition and implementation of the preliminary evaluation unit, the system can effectively perform priority assessment on the received data to ensure that urgent and important issues can be handled quickly and appropriately. The application of this method not only improves the efficiency of data processing, but also strengthens the kindergarten's ability to respond to potential problems, thereby greatly improving parent satisfaction and the kindergarten's operational efficiency. By accurately determining which issues need to be handled as a priority, the kindergarten can better allocate resources and attention and optimize the overall management process.

[0100] The best case library building module includes a case analysis unit, a case building unit, a case update unit and a case storage unit; wherein:

[0101] Case analysis unit: used to analyze the data received from the data analysis module, including the analysis results of the topic, sentiment and urgency of the questions, and identify the frequently occurring questions with high importance or urgency based on the analysis results;

[0102] Case Building Unit: Based on the analysis results of the case analysis unit, new cases are constructed, including the development of standard responses and the definition of best practice examples. Each example is used to address frequently occurring parent issues. Each new case will detail the background of the issue, the recommended response measures, the expected results, and the follow-up actions.

[0103] Case Update Unit: used to monitor the validity and timeliness of cases in the existing case library, regularly review the case library, and adjust or update the case content based on the latest data analysis results to ensure that the case library always reflects current parent issues and effective response strategies;

[0104] Case storage unit: The constructed or updated cases are saved in a structured database for quick retrieval and use. This unit ensures that the cases are organized and stored in a way that promotes efficient data access and stable performance. Through the detailed configuration and implementation of the above-mentioned best case library building module, the system can continuously provide the latest and most effective response strategies to kindergarten managers and staff, which not only helps to quickly solve parents' specific problems, but also significantly improves parent satisfaction and the overall service quality of the kindergarten.

[0105] The question classification module includes a classification decision unit, a priority assessment unit, and a reply suggestion generation unit; wherein:

[0106] Classification Decision Unit: Uses a rule-based algorithm to classify questions. Each new question is classified into predefined categories based on data extracted from the case library provided by the Best Case Library Building Module, including the type of question, urgency indicator, and common problem patterns. The rule-based algorithm classifies questions based on their similarity to cases in the case library. The calculation formula specifically uses cosine similarity to determine the degree of match between questions and known categories: in, is the eigenvector of the new problem, is the feature vector of the category in the case library;

[0107] Priority Assessment Unit: Assess the priority of the problem based on its urgency and impact. Specifically, a weight allocation method is used to assign a priority score to each problem based on the urgency index, emotional intensity, and scope of the problem. The specific calculation formula is:

[0108] Priority score (p) = w 1 ×I+w 2 ×E+w 3 ×S, where I is the impact scope of the problem, E is the emotional intensity, S is the urgency of the problem, and w 1 ,w 2 ,w 3 is the weight of each factor;

[0109] Response suggestion generation unit: Based on the classification results and priority assessment, standardized response suggestions are automatically generated. The response suggestions are guided by the best practices of similar questions in the case library to ensure that the answers provided are in line with the kindergarten's policies and meet the specific needs of parents. In addition, the unit can also adjust the response content to better adapt to specific situations and add personalized elements. Through the careful design and intelligent processing of the question classification module, the system can efficiently manage various types of questions from parents, quickly and accurately classify and prioritize questions, which not only improves the kindergarten's response speed and quality to parent needs, but also optimizes resource allocation and improves work efficiency.

[0110] The automatic reply generation module includes a reply template selection unit, a content customization unit, a language optimization unit, and a reply review unit; wherein:

[0111] Response template selection unit: The Jaccard similarity algorithm is used to select the template that best matches the current question from the standardized response suggestion library. The Jaccard similarity evaluates the similarity by comparing the ratio of the intersection and union of the question feature set and the template feature set. The specific calculation formula is: Among them, A represents the feature set of the question, B represents the feature set of the template, and the Jaccard similarity value ranges from 0 to 1. The higher the value, the higher the similarity between the two sets, so the most similar reply template is selected;

[0112] Content customization unit: After selecting the basic response template, customize the response according to the details and context of the specific question, making the response more personal and specific by adding a scheduled date, event name or targeted suggestions;

[0113] Language Optimization Unit: used to adjust the language and style of the reply to ensure that it is in line with the kindergarten's communication guidelines and easy for parents to understand. Language processing technology is used to optimize the tone and politeness, making the reply both professional and friendly in expression;

[0114] Response review unit: Before the response is sent, the generated response content is subject to a final review. During the review process, the accuracy, completeness and sensitivity of the response are checked to ensure that all responses can accurately convey the required information and comply with the kindergarten's policies and parents' expectations. Through the careful design and implementation of the automatic response generation module, the system can provide a highly customized and responsive communication solution, which not only significantly improves the efficiency of teachers' response to parents' questions, but also ensures the quality and appropriateness of the response, thereby enhancing parents' satisfaction and trust. The customized response content is more in line with the actual needs of parents, and the optimized language and style make communication smoother, which helps to maintain and enhance the professional image of the kindergarten.

[0115] The public opinion monitoring module includes a data collection unit, a real-time analysis unit, a trend analysis unit, and a report generation unit; among which:

[0116] Data collection unit: used to continuously monitor and collect public feedback from social media and other online platforms. Specifically, APIs and web crawler technology are used to automatically capture posts and comments mentioning kindergartens from platforms such as Facebook, Twitter, Instagram, and online forums. The data collection strategy is regularly updated to adapt to changes in data interfaces and privacy policies of different platforms;

[0117] Real-time analysis unit: Apply stream data processing technology, including Apache Kafka or Apache Storm, to analyze the collected data. The real-time analysis tasks include sentiment analysis and core word recognition of the text. The sentiment score is calculated using the following formula: Among them, sentiment value i is the numerical value of the sentiment tag extracted from a single post, and weight i is the weight determined based on word frequency and contextual importance;

[0118] Trend analysis unit: Use time series analysis techniques, such as the ARIMA model, to analyze the changing trends of public opinion dynamics. By calculating the average and standard deviation of sentiment scores and mention frequencies within a predetermined time window, it is used to identify changing trends and abnormal fluctuations in public opinion. The specific formula is:

[0119] Where T is the time window length, and sentiment score t is the sentiment score at time t;

[0120] Report generation unit: Based on the analysis results, a comprehensive report on the kindergarten's public image and public opinion status is generated. The report includes a graphical display of indicators, a summary of focus points and suggested measures to support the kindergarten management in making information-based decisions. Through the highly automated and intelligent design of the public opinion monitoring module, the system can effectively monitor and analyze the kindergarten's public image and public opinion dynamics on social media and online platforms. The real-time and trend analysis functions help the kindergarten to understand and respond to public feedback in a timely manner, so that it can take quick measures to maintain or improve its public image. In addition, the generated report provides the kindergarten with valuable data insights and operational suggestions, enhancing the data-driven nature and effectiveness of management decisions.

[0121] The satisfaction evaluation module includes a data collection unit, a satisfaction analysis unit, a trend monitoring unit and a report generation unit; wherein:

[0122] Data collection unit: used to collect parents’ feedback and interaction data with the kindergarten, including comments, ratings and direct feedback from parents on different platforms (such as the kindergarten’s official website, social media and email). This unit also uses APIs and web crawler technology to automatically collect data and ensure the integrity and real-time nature of the data;

[0123] Satisfaction Analysis Unit: Statistical analysis techniques and machine learning algorithms are used to evaluate parents’ satisfaction. Specifically, the following calculation formula is used to analyze the sentiment tendency and satisfaction score of parents’ feedback:

[0124] Satisfaction score = α × average sentiment score + β × positive feedback ratio + γ × engagement index, where α, β, and γ are adjustment parameters optimized based on historical data to balance the impact of various factors; the average sentiment score is based on sentiment analysis results, the positive feedback ratio is the ratio of the number of positive feedback to the total number of feedback, and the engagement index is calculated based on the frequency and diversity of parent interactions;

[0125] Trend monitoring unit: Use time series analysis to monitor the long-term trend and seasonal changes of satisfaction. Specifically, the preset autoregressive moving average model (ARMA) is used to predict future satisfaction trends and provide forward-looking data support for kindergarten management. The specific formula is:

[0126] in, and θ are model parameters, ∈ t is the error term, representing random fluctuations;

[0127] Report generation unit: Based on the analysis results, a detailed satisfaction report is generated, including indicator charts, trend analysis results and improvement suggestions. The satisfaction report provides decision-making support for kindergarten management, helping them understand the dynamic changes in parent satisfaction and formulate corresponding strategies. Through the careful design and implementation of the satisfaction evaluation module, it can provide in-depth parent satisfaction analysis and trend monitoring, enabling kindergartens to optimize educational services based on data-driven insights. This systematic satisfaction evaluation not only enhances the kindergarten's ability to respond to parent needs, but also improves the overall satisfaction of parents, thereby helping to improve the kindergarten's educational quality and service level.

[0128] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The kindergarten management intelligent system driven by public opinion big data is characterized by: It includes data collection module, data analysis module, best case library construction module, question classification module, automatic reply generation module, public opinion monitoring module and satisfaction evaluation module; among which: Data collection module: used to collect interaction data between parents and kindergartens, including emails, text messages and social media messages, and transmit the data to the central database for unified management; Data analysis module: connects to the central database and applies preset natural language processing technology to analyze the collected data, extracting the topic, sentiment and information points of the problem to evaluate the importance and urgency of the problem; Best Case Library Building Module: Based on the output of the data analysis module, build and update the case library containing standard responses and best practice examples; Question Classification Module: Classify and prioritize questions based on the case library provided by the Best Case Library Building Module, and generate standardized response suggestions; Automatic reply generation module: Generates customized reply content based on the standardized reply suggestions provided by the question classification module to support effective communication between teachers and parents; Public opinion monitoring module: real-time monitoring and analysis of public feedback from social media and online platforms to assess the kindergarten’s public image and public opinion dynamics; Satisfaction evaluation module: Based on parents’ feedback and interaction results, evaluate parent satisfaction and provide data support for improving educational services.

2. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The data collection module includes a receiving unit, a preprocessing unit, a classification unit and a transmission unit; wherein: Receiving unit: adopts a multi-protocol interface design to be compatible with different communication platforms and formats, and is used to receive interactive data between parents and kindergartens through email, SMS and social media channels. The multi-protocol interface includes SMTP protocol for email reception, SMS gateway for SMS transmission, and social media APIs for capturing social media exchanges. The multi-protocol interface can automatically identify and receive data from different electronic communication tools and cache them; Preprocessing unit: Perform data format standardization and cleaning operations through preset automated tools, including removing invalid and irrelevant information, correcting format errors, and unifying data coding; Classification unit: Use preset rules to preliminarily classify the preprocessed data for subsequent processing. The specific classification is based on the basic attributes of the data type and content. The classification automatically labels each piece of data through the classification algorithm; Transmission unit: Encryption technology and data transmission protocol are used to transmit classified and pre-processed data to the central database. The specific technologies used include TLS / SSL encryption protocol to ensure the security and integrity of data during transmission and prevent data from being illegally accessed or tampered with during transmission.

3. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The data analysis module includes a connection unit, a natural language processing unit and a preliminary evaluation unit; wherein: Connection unit: Uses standard database connection protocols to establish a stable connection with the central database, providing instant data flow for natural language processing and subsequent analysis; Natural language processing unit: configures preset natural language processing technology to parse and understand text data transmitted from the data collection module. Specific technologies include semantic analysis, sentiment analysis, and core word extraction to extract the topic, sentiment, and information points of the question from the text; Preliminary evaluation unit: Use decision tree technology to perform preliminary importance and urgency evaluation on the data obtained from the natural language processing unit to screen out the data to be processed in priority and mark the screened data as high priority.

4. The kindergarten management intelligent system driven by public opinion big data according to claim 3 is characterized in that: The natural language processing unit comprises: Text preprocessing: First, the received text data is cleaned and standardized, including removing useless characters, punctuation marks, and converting all characters to lowercase, in preparation for more in-depth language processing, ensuring the consistency and cleanliness of the text data, and laying the foundation for subsequent analysis; Semantic analysis technology: The bag-of-words model and TF-IDF weight calculation are used to identify core words and phrases in the text. The calculation formula of TF-IDF is: TF-IDF(t, d) = TF(t, d) × IDF(t), where TF(t, d) is the frequency of occurrence of term t in document d, and IDF(t) is the inverse document frequency. The calculation method is: is the total number of documents, and the denominator is the number of documents containing the term; Sentiment analysis: Use VADER's sentiment analysis algorithm to assess the sentiment tendency of the text. VADER is a sentiment analysis tool based on dictionaries and rules, which is used for social media texts. It associates each word in the text with a sentiment score and calculates the sentiment score of the entire text. The sentiment score formula is: Sentiment score = ∑ (term score) × modifier, where the modifier is adjusted according to the negative words and emphasis words before the word to enhance the accuracy of the analysis; Information point extraction: Combine natural language processing technology and named entity recognition methods to identify and extract specific information points from the text, including names, places, and dates.

5. The kindergarten management intelligent system driven by public opinion big data according to claim 4 is characterized in that: The initial assessment unit includes: Decision tree construction: First, a decision tree model is built based on historical data and expert knowledge. The model aims to evaluate the importance and urgency of the question based on the characteristics of the input data, the question topic, sentiment score, and information points obtained from the natural language processing unit; Feature extraction: Extracting features required by the decision tree model from the data received by the natural language processing unit. Specific features include topic relevance score, sentiment tendency, and urgency index; Decision tree evaluation algorithm: Apply the decision tree algorithm to evaluate the extracted features and output the importance and urgency rating of each problem. The evaluation process formula of the decision tree is: Y = f(X), where X is the vector of input features and Y is the output rating, indicating the priority of problem handling; Priority marking step: Based on the ratings output by the decision tree, the data is automatically marked into different priority categories, and high priority issues will be marked as urgent.

6. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The best case library construction module includes a case analysis unit, a case construction unit, a case update unit and a case storage unit; wherein: Case analysis unit: used to analyze the data received from the data analysis module, including the analysis results of the topic, sentiment and urgency of the questions, and identify the frequently occurring questions with high importance or urgency based on the analysis results; Case Building Unit: Based on the analysis results of the case analysis unit, new cases are constructed, including the development of standard responses and the definition of best practice examples. Each example is used to address frequently occurring parent issues. Each new case will detail the background of the issue, the recommended response measures, the expected results, and the follow-up actions. Case Update Unit: used to monitor the validity and timeliness of cases in the existing case library, regularly review the case library, and adjust or update the case content based on the latest data analysis results to ensure that the case library always reflects current parent issues and effective response strategies; Case Storage Unit: Saves constructed or updated cases in a structured database for quick retrieval and use.

7. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The question classification module includes a classification decision unit, a priority assessment unit and a reply suggestion generation unit; wherein: Classification Decision Unit: Uses a rule-based algorithm to classify questions. Each new question is classified into predefined categories based on data extracted from the case library provided by the Best Case Library Building Module, including the type of question, urgency indicator, and common problem patterns. The rule-based algorithm classifies questions based on their similarity to cases in the case library. The calculation formula specifically uses cosine similarity to determine the degree of match between questions and known categories: Similarity in, is the eigenvector of the new problem, is the feature vector of the category in the case library; Priority Assessment Unit: Assess the priority of the problem based on its urgency and impact. Specifically, a weight allocation method is used to assign a priority score to each problem based on the urgency index, emotional intensity, and scope of the problem. The specific calculation formula is: Priority score (p) = w1×I+w2×E+w3×S, where I is the impact scope of the problem, E is the emotional intensity, S is the urgency of the problem, and w1, w2, and w3 are the weights of each factor; Response suggestion generation unit: Based on the classification results and priority assessment, standardized response suggestions are automatically generated. The response suggestions are guided by the best practices of similar questions in the case library to ensure that the answers provided are in line with the kindergarten's policies and meet the specific needs of parents.

8. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The automatic reply generation module includes a reply template selection unit, a content customization unit, a language optimization unit and a reply review unit; wherein: Response template selection unit: The Jaccard similarity algorithm is used to select the template that best matches the current question from the standardized response suggestion library. The Jaccard similarity evaluates the similarity by comparing the ratio of the intersection and union of the question feature set and the template feature set. The specific calculation formula is: Among them, A represents the feature set of the question, B represents the feature set of the template, and the Jaccard similarity value ranges from 0 to 1. The higher the value, the higher the similarity between the two sets, so the most similar reply template is selected; Content customization unit: After selecting the basic response template, customize the response according to the details and context of the specific question, making the response more personal and specific by adding a scheduled date, event name or targeted suggestions; Language Optimization Unit: used to adjust the language and style of responses to ensure that they comply with the kindergarten's communication guidelines and are easy for parents to understand. Language processing technology is used to optimize the tone and politeness. Response Review Unit: Before the response is sent, the generated response content is subject to final review. During the review process, the accuracy, completeness and sensitivity of the response are checked to ensure that all responses can accurately convey the required information.

9. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The public opinion monitoring module includes a data collection unit, a real-time analysis unit, a trend analysis unit and a report generation unit; wherein: Data collection unit: used to continuously monitor and collect public feedback from social media and other online platforms, specifically using APIs and web crawler technology to automatically capture posts and comments mentioning kindergartens from platforms and online forums; Real-time analysis unit: Apply stream data processing technology, including Apache Kafka or Apache Storm, to analyze the collected data. The real-time analysis tasks include sentiment analysis and core word recognition of the text. The sentiment score is calculated using the following formula: Sentiment Value i ×Weight i , where sentiment value i is the numerical value of the sentiment tag extracted from a single post, and weight i is the weight determined based on word frequency and contextual importance; Trend analysis unit: Use time series analysis technology to analyze the changing trends of public opinion dynamics. By calculating the average and standard deviation of sentiment scores and mention frequencies within a predetermined time window, it is used to identify changing trends and abnormal fluctuations in public opinion. The specific formula is: Sentiment score t , where T is the time window length, and sentiment score t is the sentiment score at time t; Report generation unit: Based on the analysis results, a comprehensive report on the kindergarten’s public image and public opinion status is generated. The report includes a graphical display of indicators, a summary of focus points and suggested measures to support the kindergarten management in making information-based decisions.

10. The kindergarten management intelligent system driven by public opinion big data according to claim 1 is characterized in that: The satisfaction evaluation module includes a data collection unit, a satisfaction analysis unit, a trend monitoring unit and a report generation unit; wherein: Data collection unit: used to collect parents’ feedback and interaction data with the kindergarten, including parents’ comments, ratings and direct feedback on different platforms; Satisfaction Analysis Unit: Statistical analysis techniques and machine learning algorithms are used to evaluate parents’ satisfaction. Specifically, the following calculation formula is used to analyze the sentiment tendency and satisfaction score of parents’ feedback: Satisfaction score = αβ×average sentiment score + β×percentage of positive feedback + γ×engagement index, where α, β, and γ are adjustment parameters optimized based on historical data to balance the impact of various factors; the average sentiment score is based on sentiment analysis results, the positive feedback percentage is the ratio of the number of positive feedbacks to the total number of feedbacks, and the engagement index is calculated based on the frequency and diversity of parent interactions; Trend monitoring unit: Use time series analysis to monitor the long-term trend and seasonal changes of satisfaction. Specifically, use the preset autoregressive moving average model to predict future satisfaction trends and provide forward-looking data support for kindergarten management. The specific formula is: in, and θ are model parameters, ∈ t is the error term, representing random fluctuations; Report generation unit: Generates detailed satisfaction reports based on the analysis results, including indicator charts, trend analysis results and improvement suggestions. The satisfaction report provides decision support for kindergarten management.

Images