A digital consulting management method and system for engineering information based on the Internet of Things

Abstract

The present invention discloses a method and system for digital consulting management of engineering information based on the Internet of Things, and relates to the field of digital information consulting. The method comprises the following steps: S1, obtaining a consultant information group and a consulting question information group; S2, obtaining a consultant authority group and performing reply authority judgment; S3, classifying the consulting question information and then sorting the consulting question information; S4, generating a reply to the consulting question information, and calculating the problem association value of the consulting question information; S5, obtaining a recommended question parameter, and outputting a reply to the consulting question information, a recommended question parameter, and a recommended question reply; S6, collecting and optimizing the reply to the consulting question information, the recommended question parameter, and the recommended question reply of the consultant. The present invention reduces the need for manual operation and improves the efficiency and speed of data processing by automatically acquiring, cleaning, and classifying consulting information.

Description

Technical Field

[0001] The present invention relates to the field of information digitalization consultation, and in particular to an engineering information digitalization consultation management method and system based on the Internet of Things. Background Art

[0002] The Internet of Things is a technical concept covering a wide range of fields. It refers to a system that connects various physical devices through the Internet. The Internet of Things enables objects in the physical world to be remotely sensed and controlled through the Internet, thereby realizing intelligent operation and management. The core advantage of the Internet of Things is that it can interconnect objects in the physical world, collect data, and make intelligent decisions based on this data, thereby improving efficiency and convenience. It can also open the door to innovation, create new services and applications, and improve people's quality of life.

[0003] Digital consulting management of engineering information plays a vital role in modern engineering project management. It combines digital technology with traditional engineering consulting services to provide a more efficient, accurate and dynamic management method. Through digital management tools and platforms, the collection, storage, processing and analysis of engineering information become faster and more accurate, reducing errors and delays in manual processing and improving the efficiency of the entire project management process.

[0004] However, the existing digital consulting management methods for engineering information do not consider the combination of consulting issues and the authority of consultants when using them, which makes it impossible to effectively restore engineering information based on the authority of consultants during consultation, resulting in data leakage during digital consulting of engineering information, affecting the data security of engineering information. At the same time, no related information is recommended for the consulting issues, resulting in a suboptimal user experience for consultants when consulting on issues, which greatly affects the efficiency of the use of digital consulting management methods for engineering information.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related technology, the present invention proposes a method and system for digital consulting management of engineering information based on the Internet of Things to overcome the above-mentioned technical problems existing in the existing related technology.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, there is provided a method for digital consulting management of engineering information based on the Internet of Things, comprising the following steps:

[0009] S1. Obtain the consultant information group and the consulting question information group, and clean the consultant information group and the consulting question information group;

[0010] S2. Analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result;

[0011] S3. Classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the classification results of the consulting question information;

[0012] S4. Based on the results of the consultation question information sorting, generate a consultation question information reply through multi-dimensional analysis, extract consultation feature parameters of the consultation question information, and calculate the question relevance value of the consultation question information based on the consultation feature parameters;

[0013] S5. Setting a predicted question database and an association matching rule, matching the question association value with the predicted question database according to the association matching rule, obtaining recommended question parameters, and configuring recommended question replies based on the recommended question parameters, and outputting consultation question information replies, recommended question parameters, and recommended question replies;

[0014] S6. Collect the consultants’ responses to consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question responses, recommended question parameters and recommended question responses based on the evaluation results.

[0015] As a preferred solution, the internal consulting question information of the consulting question information group is classified according to the reply authority judgment result, the classification weighting rule is preset, the classification result of the consulting question information is weighted based on the classification weighting rule, and the consulting question information is sorted based on the weighting result of the consulting question information classification result, including the following steps:

[0016] S31, defining a consulting question information classification standard, and setting a consulting question information classification rule based on the consulting question information classification standard, and classifying the consulting question information according to the consulting question information classification rule;

[0017] S32, extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting a problem characteristic analysis strategy, and obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy;

[0018] S33, presetting classification weighting rules, and weighting the classification results of the consulting question information through the classification weighting rules and question characteristic parameters;

[0019] S34. Based on the weighted results of the consulting question information classification results, the consulting question information is sorted in descending order by sorting by the maximum value to obtain the consulting question information sorting results.

[0020] As a preferred solution, extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting a problem characteristic analysis strategy, obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy includes the following steps:

[0021] S321, defining a feature parameter extraction rule, and obtaining feature parameters of the consulting question information based on the feature parameter extraction rule;

[0022] S322, defining the target of the problem characteristic parameter, and analyzing the correlation parameters between the characteristic parameters of the consulting question information and the target of the problem characteristic parameter;

[0023] S323. Setting a problem characteristic analysis strategy through mapping based on the associated parameters, and verifying and optimizing the problem characteristic analysis strategy.

[0024] As a preferred solution, presetting classification weighting rules and weighting the classification results of consulting question information by using the classification weighting rules and question characteristic parameters include the following steps:

[0025] S331, defining classification weighting targets and classification weight ratio rules, and extracting weighting parameters from problem characteristic parameters based on the classification weighting targets;

[0026] S332, setting classification weighting rules according to weighting parameters, and using classification weighting rules and classification weight ratio rules to weight the consulting question information within the consulting question information classification results;

[0027] S333. Calculate the weight parameters of the consulting question information according to the results of weighting the consulting question information, and verify and optimize the weight parameters.

[0028] As a preferred solution, based on the results of the consultation question information sorting, a consultation question information reply is generated through multi-dimensional analysis, and consultation feature parameters of the consultation question information are extracted. The question association value of the consultation question information is calculated based on the consultation feature parameters, including the following steps:

[0029] S41, according to the result of sorting the consulting question information, obtaining the consulting question characteristics of the consulting question information through dimensional analysis in turn, and formulating the consulting question information reply based on the consulting question characteristics;

[0030] S42, presetting consultation feature extraction rules, and extracting consultation feature parameters of consultation question information through the consultation feature extraction rules and natural language processing algorithms;

[0031] S43, according to the consultation characteristic parameters, calculate the relevance value of the consultation question information through the relevance algorithm, preset the relevance value division threshold, compare the relevance value division threshold with the relevance value of the consultation question information, and obtain the relevance between the consultation question information;

[0032] S44. Verify and optimize the correlation between the consulting question information, and output the correlation value and correlation status of the consulting question information.

[0033] As a preferred solution, presetting consultation feature extraction rules, and extracting consultation feature parameters of consultation question information through consultation feature extraction rules and natural language processing algorithms include the following steps:

[0034] S421, analyzing the consulting question information, defining a feature extraction target for the consulting question information, and setting consulting feature extraction rules based on the feature extraction target;

[0035] S422, extracting words from the consulting question information according to the consulting feature extraction rule, and performing part-of-speech tagging on the words;

[0036] S423, using a support vector machine algorithm to build a part-of-speech analysis model, and using the part-of-speech analysis model to perform sentiment analysis on the words to obtain sentiment annotations;

[0037] S424: Generate consultation feature parameters according to the sentiment analysis results and word part-of-speech tagging results.

[0038] As a preferred solution, the expression of the part-of-speech analysis model is:

[0039] ;

[0040] in, Calculate results for sentiment analysis of words;

[0041] K is the kernel function;

[0042] a is a training dataset containing historical part-of-speech tags and historical sentiment tags for words;

[0043] x is the part-of-speech tag and sentiment tag of the input word;

[0044] B is the bias term;

[0045] T is the transpose operation;

[0046] is the weight vector of the input word.

[0047] As a preferred solution, according to the consultation feature parameters, the association value of the consultation question information is calculated by the association algorithm, and the association value division threshold is preset, and the association value division threshold is compared with the association value of the consultation question information. The acquisition of the association between the consultation question information includes the following steps:

[0048] S431, setting a relevance algorithm library and an algorithm matching strategy, matching the consultation feature parameters with the relevance algorithm inside the relevance algorithm library through the algorithm matching strategy, and using the matching relevance algorithm to calculate the relevance value of the consultation question information;

[0049] S432, defining an association value division target, and setting an association value division threshold based on the association value division target;

[0050] S433. Compare the correlation value division threshold with the consulting question information correlation value, and perform verification analysis on the comparison result to obtain the correlation between the consulting question information.

[0051] As a preferred solution, a relevance algorithm library and an algorithm matching strategy are set up, and the consultation feature parameters are matched with the relevance algorithm in the relevance algorithm library through the algorithm matching strategy, and the relevance value of the consultation question information is calculated by using the matching relevance algorithm, including the following steps:

[0052] S4311. Set up a correlation algorithm library, define algorithm matching standards, and set up an algorithm matching strategy based on the algorithm matching standards;

[0053] S4312, according to the consultation feature parameters, obtaining feature quantization parameters of the consultation feature parameters through feature conversion;

[0054] S4313. Match the feature quantization parameters with the correlation algorithm library through the algorithm matching strategy, obtain the correlation algorithm, use the matching correlation algorithm to calculate the correlation value of the consulting question information, and verify the obtained correlation value.

[0055] According to another aspect of the present invention, there is provided an engineering information digital consulting management system based on the Internet of Things, the system comprising:

[0056] An information acquisition module is used to acquire a consultant information group and a consulting question information group, and clean the consultant information group and the consulting question information group;

[0057] The information judgment module is used to analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result;

[0058] A weighting and sorting module is used to classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the consulting question information classification results;

[0059] A correlation calculation module is used to generate a consultation question information reply through multi-dimensional analysis based on the consultation question information sorting result, extract consultation feature parameters of the consultation question information, and calculate the question correlation value of the consultation question information based on the consultation feature parameters;

[0060] The recommendation output module is used to set the predicted question database and the association matching rules, match the question association value with the predicted question database according to the association matching rules, obtain the recommended question parameters, and configure the recommended question replies based on the recommended question parameters, and output the consultation question information replies, recommended question parameters and recommended question replies;

[0061] The optimization and adjustment module is used to collect the consultants' responses to consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question information responses, recommended question parameters and recommended question responses based on the evaluation results;

[0062] The information acquisition module, the information judgment module, the weighted ranking module, the association calculation module, the recommendation output module and the optimization adjustment module are connected in sequence.

[0063] The beneficial effects of the present invention are:

[0064] 1. The present invention reduces the need for manual operations by automatically acquiring, cleaning and classifying consulting information, thereby improving the efficiency and speed of data processing, enabling consultants to respond to consulting questions more quickly, thereby improving user satisfaction and service quality. By analyzing consultant information and consulting question information, combined with authority management and classification, it ensures that consulting questions can be responded to by appropriate personnel with appropriate authority, thereby improving the relevance and professionalism of consulting responses.

[0065] 2. The present invention formulates responses based on data-driven insights through detailed analysis of consulting question information and calculation of question relevance values, identifies common problems and user demand trends, optimizes consulting content, and provides more personalized and customized responses based on in-depth analysis of question characteristics and user features and calculation of question relevance values, thereby improving user experience and satisfaction.

[0066] 3. The present invention collects feedback from consultants and optimizes responses to consulting questions based on evaluation results to achieve continuous improvement of services. The quality and effectiveness of consulting services are continuously improved through an iterative process based on feedback. By setting up a predictive question database and associated matching rules, possible consulting trends and potential problems can be identified in advance, thereby taking preventive measures to reduce future consulting needs and potential problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 is a method flow chart of a method for digital consulting management of engineering information based on the Internet of Things according to an embodiment of the present invention;

[0069] Figure 2 The present invention is a system principle block diagram of an engineering information digital consulting management system based on the Internet of Things according to an embodiment of the present invention.

[0070] In the figure:

[0071] 1. Information acquisition module; 2. Information judgment module; 3. Weighting and sorting module; 4. Association calculation module; 5. Recommendation output module; 6. Optimization and adjustment module. DETAILED DESCRIPTION

[0072] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0073] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] The present invention is further described with reference to the accompanying drawings and specific embodiments. According to one embodiment of the present invention, Figure 1 As shown, the engineering information digital consulting management method based on the Internet of Things according to an embodiment of the present invention includes the following steps:

[0075] S1. Obtain the consultant information group and the consulting question information group, and clean the consultant information group and the consulting question information group;

[0076] Specifically, obtain the basic information of consultants, such as name, professional field, title, work experience, etc., from the human resources management system, consulting service platform or other relevant databases, and use automated scripts or APIs to export the consultant information from the system, collect relevant information of consulting questions from consulting service systems, customer service records, online Q&A platforms, etc., including question description, question time, questioner information, question status, etc., and use automated scripts or APIs to batch export consulting question data.

[0077] For the consultant information group and consulting question information group, it is necessary to check and delete duplicate records to ensure the uniqueness of the data and ensure that all data follows the same format and standards, such as date and time format, text encoding, etc., and for the consultant information, ensure the consistency and standardization of information such as title and professional field. For missing information, consider filling, deleting or retaining it according to the specific situation. For example, if the professional field of some consultants is missing, infer it through historical consulting records, check for errors or unreasonable values ​​in the data, and make corrections. For example, the number of years of work experience of the consultant cannot be negative. Delete information that is irrelevant to subsequent analysis and processing, such as the consultant's personal contact information is not necessary for the ability to analyze consulting problems. The data must be processed as needed, such as extracting keywords based on the description of the consulting problem, or segmenting and standardizing the consultant's skill description.

[0078] S2. Analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result;

[0079] Specifically, confirm the completeness and accuracy of the consultant information group after cleaning, and extract key features from the consultant information, such as professional field, qualifications, historical performance, specific skills, etc., and classify and rate the consultants based on the extracted features, such as classifying personnel into junior, intermediate and senior levels according to professional qualifications and experience, and determine the scope of authority of consultants at different levels, such as senior personnel can answer more complex consulting questions, and associate the extracted personnel characteristics with the corresponding authority level, such as personnel with specific professional backgrounds obtain the authority to handle related consulting issues, and classify them according to standards such as the complexity, urgency, and professionalism of the consulting issues, apply these grading standards to the consulting problem information, classify the problems into different levels, such as high, medium, and low, and then compare the grading results of the consulting problem information group with the consultant's authority group, and based on the comparison results, determine which consultants have the authority to answer consulting questions of a specific level, and then formulate a problem assignment strategy based on the authority judgment results to ensure that each problem is handled by a consultant with the corresponding authority.

[0080] S3. Classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the classification results of the consulting question information;

[0081] Specifically, the internal consulting question information of the consulting question information group is classified according to the reply authority judgment result, the classification weighting rule is preset, the classification result of the consulting question information is weighted based on the classification weighting rule, and the consulting question information is sorted based on the weighting result of the consulting question information classification result, including the following steps:

[0082] S31, defining a consulting question information classification standard, and setting a consulting question information classification rule based on the consulting question information classification standard, and classifying the consulting question information according to the consulting question information classification rule;

[0083] Specifically, analyze the common characteristics of consulting problems, such as the complexity of the problem, the required professional knowledge field, the urgency, the customer category, the problem frequency, etc., and determine the main dimensions for classification based on the characteristics of the problem. Common classification dimensions include the type of problem, such as technology, finance, service, etc., the urgency, such as high, medium, low, and professional requirements, such as professional field or knowledge level, etc. Determine the specific classification level for each dimension. For example, the urgency can be divided into three levels: "high urgency", "medium" and "low urgency". Develop clear decision logic or conditions for each classification dimension to determine which classification level the problem should be classified into. For example, the judgment of urgency is based on the response time requirement after the problem is reported. When applying logic on multiple classification dimensions, it is necessary to consider how to integrate the classification results of these dimensions to determine the final classification of the problem. The classification rules should have a certain degree of flexibility and adaptability, and can be adjusted according to actual conditions and experience.

[0084] Apply the defined classification rules to specific consulting questions and classify them according to the characteristic parameters of the questions, including automated tools such as text analysis and natural language processing algorithms to identify the key features of the questions and perform classification. For uncertain or complex cases of automatic classification results, introduce a manual review process to ensure the accuracy of the classification. Based on feedback from actual applications and problem handling results, continuously optimize the classification standards and rules, including adjusting the classification dimensions, modifying the decision logic, or introducing new classification standards.

[0085] S32, extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting a problem characteristic analysis strategy, and obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy;

[0086] Specifically, extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting the problem characteristic analysis strategy, and obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy includes the following steps:

[0087] S321, defining a feature parameter extraction rule, and obtaining feature parameters of the consulting question information based on the feature parameter extraction rule;

[0088] Specifically, understand the structure and content of consulting question information, including the subject of the question, word usage habits, questioning methods, etc., identify the elements that are crucial to understanding and answering consulting questions, such as keywords in the questions, technical terms, nouns in specific fields, etc. According to the characteristics of the consulting questions, determine which types of parameters can be used as feature parameters, such as text-type keywords, digital-type data indicators, etc., list all possible feature parameters, which should cover the main aspects and details of the consulting questions.

[0089] For text-based consulting questions, formulate text analysis rules, such as using natural language processing technology to extract keywords, phrases or semantic tags. For questions containing numbers or data, define data extraction rules, such as identifying and extracting data values, units and related data indicators. Use keyword extraction methods to identify keywords in the question text, and apply semantic analysis techniques, such as word embedding models, to understand the deeper meaning of the question text. Develop an automated feature parameter extraction process, use programming languages ​​and related libraries, and extract feature parameters from consulting question information according to defined rules. Manually review the results of automated extraction to ensure the accuracy and completeness of the feature parameters.

[0090] S322, defining the target of the problem characteristic parameter, and analyzing the correlation parameters between the characteristic parameters of the consulting question information and the target of the problem characteristic parameter;

[0091] Specifically, identify the core attributes of the problem, such as complexity, urgency, scope of impact, professional requirements, etc. For each identified core attribute, set specific evaluation criteria. For example, urgency may be defined based on the expected resolution time window, while complexity can be evaluated based on the scope and depth of required expertise.

[0092] And create a model to map the relationship between feature parameters and problem feature parameter targets. A rule-based model determines that a problem is of high complexity if it involves multiple professional terms, or a statistical-based model, such as using a machine learning algorithm to learn the relationship between feature parameters and problem characteristics from a large amount of historical data, and converting or processing the extracted feature parameters to meet the needs of the association model, including encoding text features, normalizing numerical features, etc. Using the established association model, calculate the association parameters of the problem feature parameter target from the feature parameters, such as calculating the urgency score of the problem, the complexity level, etc.

[0093] S323. Setting a problem characteristic analysis strategy through mapping based on the associated parameters, and verifying and optimizing the problem characteristic analysis strategy.

[0094] Specifically, determine the key parameters or features in the problem, map these key parameters to the corresponding analysis strategies, and formulate corresponding problem characteristic analysis strategies based on the parameter mapping, including data collection, analysis methods, model selection, etc., and verify the formulated analysis strategies to ensure that they can effectively solve the problem, including the use of simulation data or historical data for verification, and the rationality and feasibility of the strategies through expert evaluation, and optimize the problem characteristic analysis strategies based on the verification results, including adjusting parameter mapping, improving analysis methods, or redesigning strategies.

[0095] S33, presetting classification weighting rules, and weighting the classification results of the consulting question information through the classification weighting rules and question characteristic parameters;

[0096] Specifically, presetting classification weighting rules and weighting the classification results of consulting question information by using the classification weighting rules and question characteristic parameters includes the following steps:

[0097] S331, defining classification weighting targets and classification weight ratio rules, and extracting weighting parameters from problem characteristic parameters based on the classification weighting targets;

[0098] Specifically, determine the goals that need to be classified and weighted in the problem, including different aspects or characteristics of the problem, and determine it according to the specific situation, and determine the proportion of each category in the overall weighting, based on expert opinions, data analysis results or other indicators. If there are three categories, set the weight of each category to 30%, 40% and 30% respectively. According to the classification empowerment goals, determine the weighting parameters under each category, and the parameters can be indicators or characteristics related to the category, which can be a single indicator or a combination of multiple indicators. If a category is "product quality", the relevant empowerment parameters may include "product durability", "product performance", "product appearance", etc.

[0099] S332, setting classification weighting rules according to weighting parameters, and using classification weighting rules and classification weight ratio rules to weight the consulting question information within the consulting question information classification results;

[0100] Specifically, determine the weighting parameters in the problem characteristic parameters, based on the key features or indicators determined in the previous analysis, to distinguish the importance and impact of different consulting problem information, and establish classification weighting rules based on the weighting parameters. The setting of the rules should take into account the different value ranges of the weighting parameters and the degree of impact on the problem. For example, different weighting levels or scores can be set according to the level of the weighting parameters, and the classification weighting rules and classification weight proportion rules can be combined. The classification weight proportion rule determines the proportion of different categories in the overall weighting, while the classification weighting rule determines how to assign weights to each category based on the weighting parameters.

[0101] The obtained classification weighting rules are applied to the results of the consultation problem information classification, and a weight or score is assigned to each consultation problem information, and the weighted information is used for problem analysis, sorting or decision-making. For example, in the consultation problem solving process, priority is sorted according to the weighted problem information in order to more effectively allocate resources or solve higher priority problems.

[0102] S333. Calculate the weight parameters of the consulting question information according to the results of weighting the consulting question information, and verify and optimize the weight parameters.

[0103] Specifically, determine which factors affect the importance of consulting issues, including the popularity and relevance of the issue topic, the importance of the user who asked the question, the novelty and uniqueness of the issue, the sensitivity of the field involved in the issue, the time urgency of the issue, etc., and quantify the weight factors. For each determined weight factor, it is necessary to quantify its influence on the importance of the issue. Use subjective scoring methods or statistical modeling to quantify each factor, and combine the quantified weight factors according to a certain functional relationship to obtain the comprehensive weight parameters of each issue. In practical applications, it is necessary to collect a large amount of problem data as a training set, and use the above method to calculate the weight parameters of each problem. Then, based on the actual processing effect of the problem, evaluate whether the weight parameters are reasonable, and use supervised or unsupervised machine learning algorithms to optimize the weight parameters on the training set to improve the accuracy of the weight parameters.

[0104] S34. Based on the weighted results of the consulting question information classification results, the consulting question information is sorted in descending order by sorting by the maximum value to obtain the consulting question information sorting results.

[0105] Specifically, according to the previously set classification weighting rules, weights are assigned to the classification results of each consulting question information, and the allocation of weights should take into account factors such as the urgency, complexity, and scope of impact of the problem to ensure that the weight can accurately reflect the priority of the problem. If a problem has scores on multiple classification dimensions, it is necessary to calculate the comprehensive weight of these scores, which is achieved through simple weighted average, weighting and other more complex algorithms. Sorting algorithms such as quick sort and merge sort are used to sort the consulting question information in descending order according to its comprehensive weight. According to the results of the descending order, the sorting order of the consulting question information is determined, and the problem with the largest weight will be ranked first, indicating the highest priority.

[0106] Based on the sorting order, a new list of consulting question information is generated. This list is arranged from high to low according to the priority of the questions. A sorted list of consulting question information is provided for further processing, assignment or reply. The sorting process is integrated into the consulting question management system to ensure that newly received questions can be automatically weighted and sorted. Taking into account that there may be a large number of consulting questions, the performance of the sorting algorithm needs to be optimized to ensure that the sorting process is fast and efficient. Based on actual usage and user feedback, the classification weighting rules and sorting algorithms are continuously adjusted to improve the accuracy and practicality of the sorting results.

[0107] S4. Based on the results of the consultation question information sorting, generate a consultation question information reply through multi-dimensional analysis, extract consultation feature parameters of the consultation question information, and calculate the question relevance value of the consultation question information based on the consultation feature parameters;

[0108] Specifically, based on the results of the consultation question information sorting, generating a consultation question information reply through multi-dimensional analysis, extracting consultation feature parameters of the consultation question information, and calculating the question relevance value of the consultation question information based on the consultation feature parameters include the following steps:

[0109] S41, according to the result of sorting the consulting question information, obtaining the consulting question characteristics of the consulting question information through dimensional analysis in turn, and formulating the consulting question information reply based on the consulting question characteristics;

[0110] Specifically, based on the sorting results of the consulting problem information, start the analysis from the highest priority issues, fully understand the consulting issues, determine the dimensions that need to be analyzed, such as the type of problem, urgency, professional fields involved, relevant customer information, etc. For each problem, conduct a detailed analysis based on the predetermined dimensions to understand the specific background and context of the problem and the specific needs of the customer.

[0111] Based on the results of dimensional analysis, summarize the core characteristics of each consulting question, such as urgency, complexity, scope of impact, etc., evaluate the impact that the problem may have on customers or business, as well as the urgency and importance of solving the problem, identify whether there are special needs or conditions, such as whether cross-departmental collaboration is required, or whether there are specific time limits for resolution, and design specific solution strategies or plans based on the characteristics of the consulting question, including direct answers, step-by-step guidance, referral to experts, etc. When formulating responses, consider the availability of resources and the rationality of response time to ensure that the proposed solution is feasible and can be implemented within the time frame expected by the customer, and prepare detailed response content to ensure that it not only solves the problem but is also easy for the customer to understand and implement. For complex problems, consider providing step-by-step guidance or visual solutions.

[0112] S42, presetting consultation feature extraction rules, and extracting consultation feature parameters of consultation question information through the consultation feature extraction rules and natural language processing algorithms;

[0113] Specifically, presetting a consultation feature extraction rule and extracting consultation feature parameters of consultation question information through the consultation feature extraction rule and a natural language processing algorithm includes the following steps:

[0114] S421, analyzing the consulting question information, defining a feature extraction target for the consulting question information, and setting consulting feature extraction rules based on the feature extraction target;

[0115] Specifically, we need to deeply understand the specific content of the consulting question, including the background, purpose, related fields, etc. of the question, determine the key elements in the question, such as specific terms, concepts, data points or customer-specific needs, and evaluate the structure of the consulting question information. For example, whether the question has a clear classification, such as technical, financial, management, etc., and the way the question is described, such as clarity or ambiguity, how specific or broad the question is, etc., clarify why features need to be extracted, whether it is to better understand the problem, automatically classify the problem, or to provide more accurate answers, and based on the extraction purpose, set the target parameters for feature extraction, such as keywords, concept frequency, problem urgency, required level of expertise, etc.

[0116] Determine the type of features to be extracted, such as text features, numerical features, or classification labels. According to the feature extraction goal, establish specific feature extraction logic and rules. If the goal is to identify the urgency of the problem, the rules include identifying specific urgent keywords or phrases. For text-based consulting questions, apply natural language processing techniques, such as keyword extraction, topic modeling, or sentiment analysis, to automatically identify and extract relevant features. For specific types of questions, custom extraction rules are required. For example, for technical problems, extract terms and concepts related to specific technical fields.

[0117] S422, extracting words from the consulting question information according to the consulting feature extraction rule, and performing part-of-speech tagging on the words;

[0118] Specifically, collect the consulting question text that needs to be analyzed, preprocess the collected text, including removing stop words, punctuation marks and non-text elements, and use the word segmentation tool to split the preprocessed text into separate words or vocabulary. For English, it is split by spaces and punctuation marks. For Chinese, a specific word segmentation algorithm is required. According to the defined feature extraction rules, key words are extracted from the word segmentation results. The rules are based on the frequency of word occurrence, a specific keyword list or professional terms related to the consulting question.

[0119] Select a natural language processing tool to perform part-of-speech tagging on each word in the text. Input the consultation question text after word segmentation into the natural language processing tool to automatically perform part-of-speech tagging on each word, such as noun, verb, adjective, etc.

[0120] S423, using a support vector machine algorithm to build a part-of-speech analysis model, and using the part-of-speech analysis model to perform sentiment analysis on the words to obtain sentiment annotations;

[0121] Specifically, in this embodiment, a support vector machine algorithm is used to construct a part-of-speech analysis model, that is, a regression function of a nonlinear proximity support vector machine (existing technology) is used as the part-of-speech analysis model in this solution.

[0122] Among them, the regression function of the nonlinear proximity support vector machine is expressed as:

[0123]

[0124]

[0125] In the formula, K ( A ,x) represents the kernel function K Apply to training dataset A and new input data x. The role of the kernel function is to map the data from the original space to a higher-dimensional feature space so that the linear model can work effectively in this new space; e The additive term usually represents a bias term, which adds a constant offset to the model; T Represents a transpose operation, which here refers to transposing the vector or matrix of the previous calculation result; β The weight vector that represents the model. In a linear regression or support vector machine (SVM) model, the weight vector is multiplied by the data points in the feature space to make a prediction. The parameter representing the Gaussian kernel is a constant, usually called the bandwidth parameter, which is used to control the smoothness of the kernel function.

[0126] Specifically, in this embodiment, the regression function of the nonlinear proximity support vector machine is briefly transformed to obtain the part-of-speech analysis model in this solution, which is expressed as follows:

[0127]

[0128] ;

[0129] in, Calculate results for sentiment analysis of words;

[0130] K is a kernel function used to map the original feature space into a high-dimensional space, and is applied to the training data set a and the part-of-speech tag and sentiment tag x of the input word. In this embodiment, the first use of K(a,x) is to convert the text data into a form more suitable for analysis, such as converting words or phrases into word vectors. The second application of K is to further process or refine the features for better part-of-speech and sentiment analysis;

[0131] a is a training dataset containing historical part-of-speech tags and historical sentiment tags for words;

[0132] x is the part-of-speech tag and sentiment tag of the input word, that is, the new word or text segment for which the part of speech and sentiment are to be analyzed in this embodiment;

[0133] B is a bias term, which is used to adjust the data after kernel function conversion and introduce necessary offsets into the model to improve the accuracy and adaptability of the model;

[0134] T is a transpose operation, which is used to adjust the dimension of the data so that it is consistent with the weight vector β Perform correct matrix multiplication;

[0135] is the weight vector of the input word, which contains the parameters learned by the model through training. In this embodiment, β The weights associated with each part-of-speech or sentiment category are used to map the transformed features to the prediction results of part-of-speech and sentiment.

[0136] In this embodiment, the expression of the part-of-speech analysis model processes the training data and the new input data by applying the kernel function twice, and then β This approach allows the model to exploit complex patterns in high-dimensional feature spaces to improve the accuracy and efficiency of part-of-speech and sentiment analysis.

[0137] S424: Generate consultation feature parameters according to the sentiment analysis results and word part-of-speech tagging results.

[0138] Specifically, select a suitable sentiment analysis tool or library, evaluate the sentiment tendency of the text, and divide it into positive, neutral and negative according to the evaluation results, then perform sentiment analysis on the consulting question text to obtain the overall sentiment tendency score or classification, identify and extract words with important meanings according to the previous part-of-speech tagging results, such as entities represented by nouns, actions represented by verbs, attributes or emotions expressed by adjectives, etc., count the distribution of various important parts of speech, such as nouns, verbs, adjectives, etc. in consulting questions, understand the focus and emotional tendency of the questions, integrate the results of sentiment analysis and the keyword distribution of part-of-speech tagging to form a set of comprehensive feature parameters, define feature parameters based on the above analysis, such as the emotional tendency of consulting questions, the number of key nouns mentioned in the questions, the number of action indicators in the questions and the number of emotional expression words in the questions, use these feature parameters to more accurately classify consulting questions, such as giving priority to questions with strong negative emotional tendencies, and formulate personalized response strategies based on emotional tendencies and part-of-speech distribution, such as taking a more cautious and sympathetic approach to questions with strong negative emotions.

[0139] S43, according to the consultation characteristic parameters, calculate the relevance value of the consultation question information through the relevance algorithm, preset the relevance value division threshold, compare the relevance value division threshold with the relevance value of the consultation question information, and obtain the relevance between the consultation question information;

[0140] Specifically, according to the consultation feature parameters, the association value of the consultation question information is calculated by the association algorithm, and the association value division threshold is preset, and the association value division threshold is compared with the association value of the consultation question information, and the association between the consultation question information is obtained, which includes the following steps:

[0141] S431, setting a relevance algorithm library and an algorithm matching strategy, matching the consultation feature parameters with the relevance algorithm inside the relevance algorithm library through the algorithm matching strategy, and using the matching relevance algorithm to calculate the relevance value of the consultation question information;

[0142] Specifically, setting a correlation algorithm library and an algorithm matching strategy, matching the consultation feature parameters with the correlation algorithm inside the correlation algorithm library through the algorithm matching strategy, and using the matching correlation algorithm to calculate the correlation value of the consultation question information includes the following steps:

[0143] S4311. Set up a correlation algorithm library, define algorithm matching standards, and set up an algorithm matching strategy based on the algorithm matching standards;

[0144] Specifically, bring together a series of algorithms that are suitable for your data and problem types, including various machine learning algorithms, such as classification algorithms, clustering algorithms, regression algorithms, etc., implement a standard interface for each algorithm, including data input, model training, prediction output and other functions, determine the performance indicators used to evaluate the matching effect of the algorithm, such as accuracy, recall rate, F1 score, mean square error, etc., consider the characteristics of the data, such as data type, data scale, feature dimension, data distribution characteristics, etc., and different algorithms may have different requirements and applicability for data, consider business needs, such as real-time prediction, interpretability of the model, resource limitations, etc.

[0145] Perform performance tests on the algorithms in the relevance algorithm library, use defined performance indicators to evaluate the performance of the algorithms on specific data sets, and formulate algorithm matching strategies based on the algorithm performance test results and business needs, including selecting the best algorithms for different types of problems, automatically recommending algorithms based on data features, etc. Define matching rules, which recommend or select algorithms based on data features and business needs. For example, for high-dimensional sparse data, use support vector machines or principal component analysis for dimensionality reduction. Implement algorithm matching strategies in data processing and analysis processes to ensure that each problem is processed using the most appropriate algorithm. Continuously optimize the algorithm library and matching strategies based on the performance of the algorithms in actual applications and emerging algorithms, and regularly re-evaluate and adjust algorithm matching standards and strategies.

[0146] S4312, according to the consultation feature parameters, obtaining feature quantization parameters of the consultation feature parameters through feature conversion;

[0147] Specifically, remove or fill missing values, eliminate outliers, perform stemming, stop word removal, lowercase, etc. on text data, calculate the mean and standard deviation of each numerical feature, and apply it to each numerical feature, determine the maximum and minimum values ​​of each numerical feature, and then apply normalization to each numerical feature, determine the categorical features in the data set, perform one-hot encoding on the categorical features, convert each category into a binary vector, create new interactive features based on existing features, create polynomial combinations of features, such as the square and cube of features, analyze the variance of each feature, remove those features with very small variance values, calculate the correlation between features and the target variable, retain those features that are highly correlated with the target variable, use packaging methods such as recursive feature elimination or embedding to select important features, if there are a large number of features, use dimensionality reduction techniques to reduce the number of features while retaining the original information as much as possible, for visualization or specific types of classification tasks, use t-SNE or linear discriminant analysis for dimensionality reduction, split the data set into training and test sets to evaluate model performance, and use cross-validation to more accurately evaluate model performance.

[0148] S4313. Match the feature quantization parameters with the correlation algorithm library through the algorithm matching strategy, obtain the correlation algorithm, use the matching correlation algorithm to calculate the correlation value of the consulting question information, and verify the obtained correlation value.

[0149] Specifically, ensure that the consulting feature parameters have been successfully quantified through the above-mentioned feature conversion process, including standardization, normalization, encoding and other steps, to form quantitative feature parameters suitable for algorithm processing, define indicators used to evaluate algorithm performance, such as accuracy, recall rate, AUC value, etc., define the criteria for algorithm selection based on the nature of the problem and the characteristics of the data, and design experiments to evaluate the performance of different algorithms on given feature quantification parameters, including cross-validation, test set evaluation and other methods.

[0150] Organize a library of various algorithms, including Pearson correlation coefficient, Spearman rank correlation, cosine similarity, etc. Evaluate and match the algorithms in the correlation algorithm library according to the algorithm matching strategy and defined performance evaluation indicators, run multiple experiments, compare the performance of different algorithms on the same data set, and based on the experimental results, select the correlation algorithm with the best performance or the most suitable for the current data characteristics and business needs. Use the selected correlation algorithm to process the feature quantification parameters in the consulting question information and calculate the correlation between different consulting questions.

[0151] Analyze the calculated correlation, identify highly relevant consulting questions, and use them in various application scenarios such as question classification, recommendation system, priority allocation, etc., and then verify the correlation value calculated by the algorithm through actual cases or known correlations. If two questions are known to be related, verify whether the correlation value of this pair of questions is significantly higher than that of a random pair of questions. Evaluate the accuracy and reliability of the algorithm based on the verification results. If the algorithm performs poorly, it may be necessary to reselect the algorithm or adjust the matching strategy.

[0152] S432, defining an association value division target, and setting an association value division threshold based on the association value division target;

[0153] Specifically, understand business goals and requirements. The goals of the association value division should be closely related to the business goals, such as improving the accuracy of the recommendation system, discovering related products in sales data, or identifying related topics in text data. Evaluate the importance of different degrees of correlation to business decisions. For example, in a product recommendation system, a very strong association between products means a high probability of co-purchase, while a weaker association may have no actual marketing value. Analyze known association instances in historical data and the impact of these associations on business goals, and consult relevant research or case studies to understand the association value division criteria and thresholds commonly used in similar situations.

[0154] Use statistical analysis methods to set preliminary thresholds for association values, and combine the opinions and experience of domain experts to set or adjust thresholds. Then verify the effectiveness of the set thresholds through actual experiments or simulation tests. For example, use A / B testing in recommendation systems to evaluate the recommendation effects under different thresholds. Refine and adjust the thresholds based on test results and feedback from actual applications.

[0155] S433. Compare the correlation value division threshold with the consulting question information correlation value, and perform verification analysis on the comparison result to obtain the correlation between the consulting question information.

[0156] Specifically, the value calculated by the correlation algorithm reflects the strength of correlation between each consulting question, and is then determined based on business needs and statistical analysis to distinguish between "relevant" and "irrelevant" or different levels of correlation. For each pair of consulting questions, the correlation value is compared with the set threshold. If the correlation value is higher than the threshold, the two questions are considered to be related. If it is lower than the threshold, they are considered to be irrelevant. The comparison results of each pair of consulting questions are recorded and marked as whether they are considered to be related or irrelevant. Some samples are randomly selected from the comparison results, including examples judged to be related and irrelevant, and their actual correlation is manually checked to verify the accuracy of the algorithm's judgment. Statistical methods, such as confusion matrix, precision, recall rate and other indicators are used to evaluate the overall accuracy and reliability of the comparison results. The verification results are analyzed to identify whether specific types of questions are systematically misclassified, indicating that the correlation algorithm or threshold needs to be adjusted.

[0157] S44. Verify and optimize the correlation between the consulting question information, and output the correlation value and correlation status of the consulting question information.

[0158] S5. Setting a predicted question database and an association matching rule, matching the question association value with the predicted question database according to the association matching rule, obtaining recommended question parameters, and configuring recommended question replies based on the recommended question parameters, and outputting consultation question information replies, recommended question parameters, and recommended question replies;

[0159] Specifically, collect and build a predictive question database, which includes possible consulting questions and their responses, compiled according to historical consulting records or expert knowledge, and perform characterization on each question in the database, such as extracting keywords, sentiment analysis, part-of-speech tagging, etc., to facilitate subsequent matching work, and define a set of association matching rules based on question characteristics and business needs, including keyword matching, sentiment similarity, topic consistency, etc., and select or develop algorithms suitable for executing these matching rules, such as text similarity algorithms, machine learning classifiers, etc.

[0160] The input consulting question is characterized and then matched with the questions in the prediction question database according to the defined association matching rules, and the association value is calculated. Based on the association value obtained from the match, a batch of questions most relevant to the input question are screened out from the database as recommended questions. For each screened recommended question, its predefined reply is extracted, and the reply is appropriately customized and adjusted according to the specific content and context of the input question to ensure the relevance and accuracy of the reply. The reply to the input consulting question, the recommended question parameters such as question title, association value, etc. and the reply to each recommended question are integrated together, and the above integrated information is output in a clear format, including direct replies to the consulting question, a list of recommended questions and their replies, etc.

[0161] S6. Collect the consultants’ responses to consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question responses, recommended question parameters and recommended question responses based on the evaluation results.

[0162] Specifically, provide consultants with convenient feedback channels, such as online questionnaires, rating or commenting directly on the consulting platform, collecting feedback via email, etc.; clarify the key indicators for feedback collection, such as the accuracy, relevance, and comprehensibility of responses, and the relevance of recommended questions; after the consulting response, actively invite consultants to evaluate the responses and recommended questions; set up a mechanism to allow consultants to proactively provide feedback at any time, ensure that they can easily report problems encountered during use; conduct statistical analysis on the collected quantitative data such as ratings, identify low-scoring items and high-scoring items and analyze the reasons.

[0163] Conduct content analysis on text feedback, categorize issues and suggestions, extract improvement points, and then adjust the content and method of responses to consulting questions based on the feedback to ensure that responses are more accurate, comprehensive, and easy to understand. Optimize the selection and sorting logic of recommended questions to ensure that recommended questions are more relevant to consulting questions. Adjust association matching rules and algorithm parameters based on feedback results to improve matching accuracy and recommendation quality. After implementing improvement measures, continue to monitor feedback indicators and evaluate the effectiveness of improvements. Treat optimization and adjustment as a continuous process and continuously iterate and optimize responses and recommendations based on new feedback information.

[0164] According to another embodiment of the present invention, Figure 2 An engineering information digital consulting management system based on the Internet of Things is shown, the system comprising:

[0165] Information acquisition module 1, used to acquire the consultant information group and the consulting question information group, and clean the consultant information group and the consulting question information group;

[0166] Information judgment module 2 is used to analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result;

[0167] The weighting and sorting module 3 is used to classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset the classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the classification results of the consulting question information;

[0168] The association calculation module 4 is used to generate a consultation question information reply through multi-dimensional analysis based on the consultation question information sorting result, extract consultation feature parameters of the consultation question information, and calculate the question association value of the consultation question information based on the consultation feature parameters;

[0169] The recommendation output module 5 is used to set the predicted question database and the association matching rules, match the question association value with the predicted question database according to the association matching rules, obtain the recommended question parameters, and configure the recommended question reply based on the recommended question parameters, and output the consultation question information reply, the recommended question parameters and the recommended question reply;

[0170] Optimization and adjustment module 6, used to collect the consultant's responses to the consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question information responses, recommended question parameters and recommended question responses based on the evaluation results;

[0171] The information acquisition module 1, the information judgment module 2, the weighted ranking module 3, the association calculation module 4, the recommendation output module 5 and the optimization adjustment module 6 are connected in sequence.

[0172] To sum up, with the help of the above-mentioned technical scheme of the present invention, the present invention reduces the need for manual operations by automatically acquiring, cleaning and classifying consulting information, thereby improving the efficiency and speed of data processing, enabling consultants to respond to consulting questions more quickly, thereby improving user satisfaction and service quality, and by analyzing consultant information and consulting question information, combined with authority management and classification, ensuring that consulting questions can be responded to by appropriate personnel with appropriate authority, thereby improving the relevance and professionalism of consulting responses.

[0173] In addition, the present invention formulates responses based on data-driven insights through detailed analysis of consulting question information and calculation of question relevance values, identifies common problems and user demand trends, optimizes consulting content and strategies, and provides more personalized and customized responses based on in-depth analysis of question characteristics and user features and calculation of question relevance values, thereby improving user experience and satisfaction.

[0174] In addition, the present invention achieves continuous improvement of services by collecting feedback from consultants and optimizing responses to consulting question information based on evaluation results. The quality and effectiveness of consulting services are continuously improved through an iterative process based on feedback. Possible consulting trends and potential problems can be identified in advance by setting up a predictive question database and associated matching rules, thereby taking preventive measures to reduce future consulting needs and potential problems.

[0175] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A digital consulting management method for engineering information based on the Internet of Things, characterized in that: The following steps are involved: S1. Obtain the consultant information group and the consulting question information group, and clean the consultant information group and the consulting question information group; S2. Analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result; S3. Classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the classification results of the consulting question information; S4. Based on the results of the consultation question information sorting, generate a consultation question information reply through multi-dimensional analysis, extract consultation feature parameters of the consultation question information, and calculate the question relevance value of the consultation question information based on the consultation feature parameters; S5. Setting a predicted question database and an association matching rule, matching the question association value with the predicted question database according to the association matching rule, obtaining recommended question parameters, and configuring recommended question replies based on the recommended question parameters, and outputting consultation question information replies, recommended question parameters, and recommended question replies; S6. Collect the consultants' responses to consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question responses, recommended question parameters and recommended question responses based on the evaluation results; The method of classifying the internal consulting question information of the consulting question information group according to the reply authority judgment result, presetting the classification weighting rule, weighting the consulting question information classification result based on the classification weighting rule, and sorting the consulting question information based on the weighting result of the consulting question information classification result includes the following steps: S31, defining a consulting question information classification standard, and setting a consulting question information classification rule based on the consulting question information classification standard, and classifying the consulting question information according to the consulting question information classification rule; S32, extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting a problem characteristic analysis strategy, and obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy; S33, presetting classification weighting rules, and weighting the classification results of the consulting question information through the classification weighting rules and question characteristic parameters; S34, based on the weighted results of the consulting question information classification results, sorting in descending order by sorting by maximum value to obtain the consulting question information sorting results; The step of extracting characteristic parameters of the consulting question information in the consulting question information classification result, obtaining the problem characteristic parameters, and setting a problem characteristic analysis strategy, and obtaining the problem characteristic parameters of the problem characteristic parameters through the problem characteristic analysis strategy includes the following steps: S321, defining a feature parameter extraction rule, and obtaining feature parameters of the consulting question information based on the feature parameter extraction rule; S322, defining the target of the problem characteristic parameter, and analyzing the correlation parameters between the characteristic parameters of the consulting question information and the target of the problem characteristic parameter; S323, setting a problem characteristic analysis strategy through mapping based on the associated parameters, and verifying and optimizing the problem characteristic analysis strategy; The method of generating a consultation question information reply through multi-dimensional analysis based on the consultation question information sorting result, extracting consultation feature parameters of the consultation question information, and calculating the question association value of the consultation question information based on the consultation feature parameters includes the following steps: S41, according to the result of sorting the consulting question information, obtaining the consulting question characteristics of the consulting question information through dimensional analysis in turn, and formulating the consulting question information reply based on the consulting question characteristics; S42, presetting consultation feature extraction rules, and extracting consultation feature parameters of consultation question information through the consultation feature extraction rules and natural language processing algorithms; S43, according to the consultation characteristic parameters, calculate the relevance value of the consultation question information through the relevance algorithm, preset the relevance value division threshold, compare the relevance value division threshold with the relevance value of the consultation question information, and obtain the relevance between the consultation question information; S44, verifying and optimizing the correlation between the consulting question information, and outputting the correlation value and correlation status of the consulting question information; The presetting of the consultation feature extraction rule and extracting the consultation feature parameters of the consultation question information by using the consultation feature extraction rule and the natural language processing algorithm comprises the following steps: S421, analyzing the consulting question information, defining a feature extraction target for the consulting question information, and setting consulting feature extraction rules based on the feature extraction target; S422, extracting words from the consulting question information according to the consulting feature extraction rule, and performing part-of-speech tagging on the words; S423, using a support vector machine algorithm to build a part-of-speech analysis model, and using the part-of-speech analysis model to perform sentiment analysis on the words to obtain sentiment annotations; S424, generating consultation feature parameters according to the sentiment analysis results and word part-of-speech tagging results; The expression of the part-of-speech analysis model is: ； in, Calculate results for sentiment analysis of words; K is the kernel function; a is a training dataset containing historical part-of-speech tags and historical sentiment tags for words; x is the part-of-speech tag and sentiment tag of the input word; B is the bias term; T is the transpose operation; is the weight vector of the input word.

2. According to the method of engineering information digital consulting management based on the Internet of Things in claim 1, it is characterized in that: The preset classification weighting rules and weighting of the classification results of the consulting question information by the classification weighting rules and question characteristic parameters include the following steps: S331, defining classification weighting targets and classification weight ratio rules, and extracting weighting parameters from problem characteristic parameters based on the classification weighting targets; S332, setting classification weighting rules according to weighting parameters, and using classification weighting rules and classification weight ratio rules to weight the consulting question information within the consulting question information classification results; S333. Calculate the weight parameters of the consulting question information according to the results of weighting the consulting question information, and verify and optimize the weight parameters.

3. According to the method of engineering information digital consulting management based on the Internet of Things in claim 1, it is characterized in that: The step of calculating the correlation value of the consulting question information by a correlation algorithm based on the consulting feature parameters, presetting a correlation value division threshold, comparing the correlation value division threshold with the correlation value of the consulting question information, and obtaining the correlation between the consulting question information includes the following steps: S431, setting a relevance algorithm library and an algorithm matching strategy, matching the consultation feature parameters with the relevance algorithm inside the relevance algorithm library through the algorithm matching strategy, and using the matching relevance algorithm to calculate the relevance value of the consultation question information; S432, defining an association value division target, and setting an association value division threshold based on the association value division target; S433. Compare the correlation value division threshold with the consulting question information correlation value, and perform verification analysis on the comparison result to obtain the correlation between the consulting question information.

4. The method for digital consulting and management of engineering information based on the Internet of Things according to claim 3 is characterized in that: The steps of setting up the relevance algorithm library and the algorithm matching strategy, matching the consultation feature parameters with the relevance algorithm in the relevance algorithm library through the algorithm matching strategy, and using the matching relevance algorithm to calculate the relevance value of the consultation question information include the following steps: S4311. Set up a correlation algorithm library, define algorithm matching standards, and set up an algorithm matching strategy based on the algorithm matching standards; S4312, according to the consultation feature parameters, obtaining feature quantization parameters of the consultation feature parameters through feature conversion; S4313. Match the feature quantization parameters with the correlation algorithm library through the algorithm matching strategy, obtain the correlation algorithm, use the matching correlation algorithm to calculate the correlation value of the consulting question information, and verify the obtained correlation value.

5. An engineering information digital consultation management system based on the Internet of Things, used to implement the engineering information digital consultation management method based on the Internet of Things as described in any one of claims 1 to 4, characterized in that: The system includes: An information acquisition module is used to acquire a consultant information group and a consulting question information group, and clean the consultant information group and the consulting question information group; The information judgment module is used to analyze the cleaned consultant information group, obtain the consultant authority group, and classify the consulting question information group, compare the consulting question information group classification result with the consultant authority group, and make a reply authority judgment based on the comparison result; A weighting and sorting module is used to classify the internal consulting question information of the consulting question information group according to the reply authority judgment result, preset classification weighting rules, weight the classification results of the consulting question information based on the classification weighting rules, and sort the consulting question information based on the weighting results of the consulting question information classification results; A correlation calculation module is used to generate a consultation question information reply through multi-dimensional analysis based on the consultation question information sorting result, extract consultation feature parameters of the consultation question information, and calculate the question correlation value of the consultation question information based on the consultation feature parameters; The recommendation output module is used to set the predicted question database and the association matching rules, match the question association value with the predicted question database according to the association matching rules, obtain the recommended question parameters, and configure the recommended question replies based on the recommended question parameters, and output the consultation question information replies, recommended question parameters and recommended question replies; The optimization and adjustment module is used to collect the consultants' responses to consulting questions, recommended question parameters and recommended question responses, and optimize and adjust the consulting question information responses, recommended question parameters and recommended question responses based on the evaluation results; The information acquisition module, the information judgment module, the weighted ranking module, the association calculation module, the recommendation output module and the optimization adjustment module are connected in sequence.

Images