Historical data-based auditing opinion generation method, device and equipment

Abstract

The invention discloses a historical data-based auditing opinion generation method, apparatus and device. The method comprises the steps of classifying historical auditing opinion data by using a completely trained auditing opinion classification model and storing the historical auditing opinion data to a corresponding tag database; determining a professional label needing to be audited from the construction drawing data to be audited, and formulating an auditing task according to the professional label; extracting auditing opinion elements from the historical auditing opinion data in a label database corresponding to the professional label; and generating an auditing opinion corresponding to the to-be-audited construction drawing data based on the auditing opinion element and the auditing task. According to the method, the historical auditing opinion data are classified and stored, so that the related auditing opinion elements are extracted from the historical auditing opinion data according to the to-be-audited construction drawing, and finally the auditing opinions corresponding to the to-be-audited construction drawing data are generated, so that the historical auditing opinion data are reasonably utilized; and the efficiency of determining the auditing opinions of the construction drawing is improved.

Description

Technical Field

[0001] The present application relates to the technical field of drawing review, and more specifically, to a method, device and equipment for generating review opinions based on historical data. Background Art

[0002] The construction of information platforms, such as the construction drawing review platform, enables centralized control and unified standards for drawing review, improving the fairness and rationality of the review. With the development of information technology, engineering drawing review has gradually realized informatization and digitization. Traditional paper drawings have been replaced by CAD (computer-aided design) drawings, which greatly improves the efficiency of drawing storage, transmission and review.

[0003] In the existing technology, manual review is still the main way to review engineering drawings. Reviewers need to be familiar with relevant regulations, laws and policy requirements, such as construction laws, building design specifications, engineering quality standards, etc. Manual review focuses on details and comprehensiveness, and can discover potential problems and hidden dangers in drawings. Reviewers need to have certain technical knowledge and professional background, and understand the design principles, structural safety, construction technology and other requirements of construction projects.

[0004] However, the audit opinion generation method in the prior art does not summarize and record special types of audit opinions during the audit process, resulting in the subsequent audit process being unable to provide good audit opinions or providing examples of erroneous audit opinions for learning and reference in subsequent construction drawing audits, which greatly reduces the efficiency of construction drawing audits and loses the potential value of historical audit opinions. Summary of the invention

[0005] In response to at least one defect or improvement need in the prior art, the present invention provides a method, device and equipment for generating audit opinions based on historical data, which is used to solve the problem that in the prior art, special types of audit opinions in the audit process are not summarized and recorded, resulting in the subsequent audit process being unable to provide good audit opinions or erroneous audit opinions for learning and reference in the subsequent construction drawing audit, which greatly reduces the efficiency of the construction drawing audit and loses the potential value of historical audit opinions.

[0006] To achieve the above object, according to a first aspect of the present invention, a method for generating audit opinions based on historical data is provided, comprising:

[0007] Use the well-trained audit opinion classification model to classify historical audit opinion data and store them in the corresponding label database;

[0008] Determine the professional labels that need to be reviewed from the construction drawing data to be reviewed, and formulate review tasks based on the professional labels;

[0009] Extracting audit opinion elements from historical audit opinion data in the label database corresponding to the professional label;

[0010] Generate audit opinions corresponding to the construction drawing data to be audited based on the audit opinion elements and audit tasks.

[0011] In a possible implementation, the historical audit opinion data is classified and stored in a corresponding label database using a well-trained audit opinion classification model, which also includes:

[0012] Obtain a fully trained audit opinion classification model;

[0013] Input the historical audit opinion data into the well-trained audit opinion classification model, and output the professional labels corresponding to the historical audit opinion data;

[0014] Establish a label database based on professional labels, and store historical review opinion data corresponding to professional labels in the label database.

[0015] In a possible implementation, the construction drawing data to be reviewed includes image data and text data; and determining the professional tags to be reviewed from the construction drawing data to be reviewed further includes:

[0016] Use graph neural network models to perform image recognition on image data and extract key factors from image data;

[0017] Perform semantic analysis on text data based on key factors to determine the professional labels of the construction drawing data to be reviewed.

[0018] In a possible implementation, the audit tasks are formulated according to professional labels, and also include:

[0019] Develop audit task templates based on historical audit opinion data;

[0020] The professional tags are matched with the audit task template to obtain the audit task.

[0021] In a possible implementation, extracting audit opinion elements from historical audit opinion data in a tag database corresponding to a professional tag also includes:

[0022] Conduct cluster analysis on historical audit opinion data to determine the basic elements of historical audit opinion data;

[0023] Build a knowledge graph based on basic elements and determine the relationship between basic elements in the knowledge graph;

[0024] Corresponding audit opinion elements are extracted from historical audit opinion data based on the association relationship and the construction drawing data to be audited.

[0025] In a possible implementation, cluster analysis is performed on the historical audit opinion data to determine the basic elements of the historical audit opinion data, and further includes:

[0026] Clean the historical audit opinion data, remove duplicate data and format them into the same data format;

[0027] Determine a vocabulary vector sequence according to the frequency of occurrence of each vocabulary in the historical review opinion data;

[0028] Based on the preset clustering algorithm, the vocabulary vector sequence is clustered to obtain the element clustering cluster;

[0029] Analyze the feature clusters to determine the basic elements represented by each feature cluster.

[0030] In a possible implementation, generating the audit opinion corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit task also includes:

[0031] Automatically fill in the preset audit opinion template based on the audit opinion elements and audit tasks to obtain the initial audit opinion;

[0032] Use natural language generation technology to adjust the initial audit opinion to obtain a transitional audit opinion;

[0033] Perform sentence detection on the transitional audit opinion, modify the detected sentence problems and obtain the corresponding audit opinion.

[0034] According to a second aspect of the present invention, there is also provided a device for generating audit opinions based on historical data, comprising:

[0035] A classification storage module, which is configured to classify the historical audit opinion data using a well-trained audit opinion classification model and store it in a corresponding label database;

[0036] A task formulation module, which is configured to determine the professional tags that need to be reviewed from the construction drawing data to be reviewed, and formulate the review tasks according to the professional tags;

[0037] An element extraction module is configured to extract audit opinion elements from historical audit opinion data in a tag database corresponding to a professional tag;

[0038] The opinion generation module is configured to generate an audit opinion corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit task.

[0039] According to a third aspect of the present invention, there is also provided an audit opinion generating device based on historical data, which includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes any one of the steps of the above-mentioned audit opinion generating method based on historical data.

[0040] According to a fourth aspect of the present invention, there is also provided a storage medium storing a computer program executable by an access authentication device, wherein when the computer program runs on the access authentication device, the access authentication device executes any of the steps of the method for generating an audit opinion based on historical data described above.

[0041] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0042] The present invention provides a method for generating audit opinions based on historical data. By classifying and storing historical audit opinions, the system can quickly find relevant historical audit opinion data according to the professional labels of the construction drawings to be audited. This method significantly reduces the time for manually reviewing a large number of historical documents, making the audit process more efficient. Using a well-trained audit opinion classification model, it can ensure that historical audit opinions are accurately classified and stored. At the same time, the audit opinions generated based on these verified historical data are more likely to cover common audit points and potential problems, thereby improving the accuracy and comprehensiveness of the audit. Through data analysis, the system can discover common problems, high-frequency problems, etc. in the audit process, and provide data support for the formulation of more scientific and reasonable audit strategies and standards. At the same time, the prediction model based on big data can also help predict potential risks. It helps to realize the automation and intelligence of the audit process, reduce manual intervention links, and reduce the probability of human errors. At the same time, through real-time monitoring and feedback mechanisms, problems in the audit process can be discovered and corrected in a timely manner, and the audit process can be continuously optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, 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 application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A flowchart of an embodiment of a method for generating audit opinions based on historical data provided by the present invention;

[0045] Figure 2 The present invention provides Figure 1 A schematic diagram of a flow chart of an embodiment of step S100;

[0046] Figure 3 The present invention provides Figure 1 A flowchart of an embodiment of step S300;

[0047] Figure 4 The present invention provides Figure 3 A flowchart of an embodiment of step S310;

[0048] Figure 5 The present invention provides Figure 1 A flowchart of an embodiment of step S400;

[0049] Figure 6 A schematic diagram of the structure of an embodiment of a device for generating audit opinions based on historical data provided by the present invention;

[0050] Figure 7 A schematic diagram of the structure of a device for generating audit opinions based on historical data provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0052] The terms "first", "second", "third", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0053] The present invention provides a method, device and equipment for generating audit opinions based on historical data, which are described below respectively.

[0054] See also Figure 1 , Figure 1 A flowchart of an embodiment of a method for generating audit opinions based on historical data provided by the present invention is provided. In a specific embodiment of the present invention, a method for generating audit opinions based on historical data is disclosed, including:

[0055] S100, using a well-trained audit opinion classification model to classify historical audit opinion data and store them in a corresponding label database;

[0056] S200, determining the professional labels that need to be reviewed from the construction drawing data to be reviewed, and formulating the review tasks according to the professional labels;

[0057] S300, extracting audit opinion elements from historical audit opinion data in a tag database corresponding to a professional tag;

[0058] S400: Generate audit opinions corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit tasks.

[0059] In the above embodiment, first, the construction drawing review opinion data of the past years are widely collected, including review records in various forms such as text descriptions, pictures, and videos, and the collected data are cleaned to remove duplicate, incomplete or irrelevant information to ensure data quality. Natural language processing (NLP) technology, such as word segmentation, part-of-speech tagging, named entity recognition, etc., are used to extract key features from the review opinions, such as problem type, severity, and involved specifications. Based on the extracted features, machine learning or deep learning algorithms (such as support vector machines, random forests, neural networks, etc.) are used to train the review opinion classification model. The model needs to be optimized through multiple iterations to ensure classification accuracy and generalization ability. The review opinions are classified according to the preset classification system (such as structural safety, fire safety, energy conservation and environmental protection, etc.) and stored in the corresponding label database. Each label database not only contains the review opinion text, but also associates specific construction drawing information, review personnel, review time and other metadata, which is convenient for subsequent query and analysis.

[0060] Using image recognition or text analysis technology, key information in the construction drawings to be reviewed can be automatically identified, such as building type, structural form, usage function, etc., so as to determine the professional tags that need to be reviewed (such as building structure, water supply and drainage, electricity, etc.). Based on the professional tags, the review tasks are intelligently formulated, including the review scope, key focus points, review standards, etc., and assigned to the corresponding reviewers or system modules.

[0061] According to the professional tags of the construction drawings to be audited, relevant historical audit opinion data is quickly retrieved from the corresponding tag database. Using NLP and data analysis technology, audit opinion elements that are highly relevant to the current audit task are extracted from historical audit opinions, such as common problems, solutions, and specification references. According to the priority and importance of the audit task, the extracted audit opinion elements are screened and sorted to ensure that the generated audit opinion is both comprehensive and highlights the key points.

[0062] Based on the specific requirements of the audit opinion elements and audit tasks, template filling, natural language generation (NLG) and other technologies are used to automatically generate audit opinions corresponding to the construction drawing data to be audited. The generated audit opinions should be clear, accurate and easy to understand. In order to improve the quality of the audit opinions, a manual review link can be set up, and experienced auditors can review and adjust the automatically generated audit opinions. The audit opinions are output to relevant personnel in the form of documents, reports or direct annotations on the construction drawings. At the same time, user feedback is collected to continuously optimize the audit opinion classification model and generation algorithm.

[0063] Compared with the prior art, the present embodiment provides a method for generating audit opinions based on historical data. By classifying and storing historical audit opinions, the system can quickly find relevant historical audit opinion data according to the professional labels of the construction drawings to be audited. This method significantly reduces the time for manually reviewing a large number of historical documents, making the audit process more efficient. Using a well-trained audit opinion classification model, it can ensure that historical audit opinions are accurately classified and stored. At the same time, the audit opinions generated based on these verified historical data are more likely to cover common audit points and potential problems, thereby improving the accuracy and comprehensiveness of the audit. Through data analysis, the system can discover common problems and high-frequency problems in the audit process, and provide data support for the formulation of more scientific and reasonable audit strategies and standards. At the same time, the prediction model based on big data can also help predict potential risks. It helps to realize the automation and intelligence of the audit process, reduce manual intervention links, and reduce the probability of human errors. At the same time, through real-time monitoring and feedback mechanisms, problems in the audit process can be discovered and corrected in a timely manner, and the audit process can be continuously optimized.

[0064] See also Figure 2 , Figure 2 The present invention provides Figure 1 In some embodiments of the present invention, the historical audit opinion data is classified and stored in the corresponding label database using the well-trained audit opinion classification model, and further includes:

[0065] S110, obtaining a fully trained audit opinion classification model;

[0066] S120, inputting the historical audit opinion data into a well-trained audit opinion classification model, and outputting professional labels corresponding to the historical audit opinion data;

[0067] S130. Establish a label database based on professional labels, and store historical review opinion data corresponding to the professional labels in the label database.

[0068] In the above embodiment, a suitable audit opinion classification model is selected according to the project requirements. This may be a model based on machine learning (such as logistic regression, decision tree, random forest) or deep learning (such as convolutional neural network CNN, recurrent neural network RNN, Transformer, etc.). Make sure that the selected model has been fully trained and its performance has been verified, which usually includes evaluating the model's accuracy, recall rate, F1 score and other indicators on an independent test set to ensure that the model can accurately classify audit opinions. Deploy the trained model to a suitable computing environment, such as a cloud server, local server or edge computing device, for subsequent data processing and classification tasks.

[0069] Before inputting the historical audit opinion data into the model, necessary data preprocessing is performed, such as noise removal, format unification, text encoding conversion, etc., to ensure that the data meets the model input requirements. In order to improve processing efficiency, a large amount of historical audit opinion data can be input into the model at one time in batch processing. At the same time, sufficient memory and computing resources are ensured during data loading and model prediction. The preprocessed historical audit opinion data is input into a well-trained audit opinion classification model. The model automatically determines and outputs the professional label corresponding to each audit opinion based on the input data. The professional labels output by the model are verified to ensure the accuracy and rationality of the classification results. For audit opinions that are misclassified or uncertain, manual review or retraining of the model can be performed to improve classification accuracy.

[0070] Design the structure of the label database according to the characteristics and requirements of professional labels, including the design of database tables, definition of fields, creation of indexes, etc., to ensure that the database can efficiently store and query a large amount of historical audit opinion data. Store each historical audit opinion data and its corresponding professional label in the label database according to the designed database structure. During the storage process, appropriate data compression and encryption technologies can be used to ensure data security and save storage space, and create indexes for key fields in the label database (such as professional labels, audit time, auditors, etc.) to improve the efficiency and accuracy of data queries. Regularly maintain the label database, including data backup, data cleaning, performance optimization, etc., to ensure the stable operation of the database and the long-term availability of data.

[0071] In some embodiments of the present invention, the construction drawing data to be reviewed includes image data and text data; determining the professional tags to be reviewed from the construction drawing data to be reviewed also includes:

[0072] Use graph neural network models to perform image recognition on image data and extract key factors from image data;

[0073] Perform semantic analysis on text data based on key factors to determine the professional labels of the construction drawing data to be reviewed.

[0074] In the above embodiment, first, the image data is preprocessed, including resizing, noise removal, contrast enhancement, etc., to improve the accuracy of the image recognition model. Convolutional neural network (CNN) or more advanced image recognition technology (such as residual network ResNet, VGGNet, etc.) is used as the front end of graph neural network (GNN) to extract high-level features such as edges, textures, shapes, etc. from the preprocessed image. The extracted image features are converted into graph structure data, where nodes can represent key objects or regions in the image, and edges reflect the spatial relationship or semantic connection between these objects or regions. The constructed graph structure data is processed using graph neural network models (such as Graph Convolutional Networks, GCN; Graph Attention Networks, GAT, etc.) to capture more complex structures and contextual information in the image. Through the output of the graph neural network, key factors in the image data are identified, such as building structure type, material usage, safety hazards, etc. These factors are crucial for the subsequent determination of professional labels.

[0075] Preprocess the text data associated with the image data, including word segmentation, part-of-speech tagging, and stop word removal, to improve the accuracy of semantic analysis. Extract keywords from the preprocessed text data, which are usually closely related to various aspects of architectural design, such as structural description, material list, design specifications, etc. Perform semantic association analysis on the extracted keywords and the key factors obtained from image recognition to find the commonalities and potential connections between the two, which helps to more comprehensively understand the content and intent of the construction drawing data. Based on the results of semantic association and combined with the preset professional label system, determine the professional labels of the construction drawing data to be reviewed. These labels may involve multiple professional fields such as structural design, water supply and drainage, electrical installation, HVAC, etc.

[0076] In some embodiments of the present invention, formulating the audit task according to the professional tag further includes:

[0077] Develop audit task templates based on historical audit opinion data;

[0078] The professional tags are matched with the audit task template to obtain the audit task.

[0079] In the above embodiment, the historical audit opinion data is deeply analyzed to identify the common audit points, concerns and potential violations under different professional labels. Based on these analysis results, one or more audit task templates are designed for each professional label. The audit task template should include but not be limited to the following aspects: audit objectives (clarifying what specific content the task is intended to verify or check), audit steps (listing in detail the specific actions or checkpoints required to complete the audit), reference standards (pointing out the specifications, standards or laws and regulations that should be followed during the audit process), and expected results (describing the audit results that should be obtained or the types of problems that should be found under normal circumstances). As the audit work progresses, feedback is continuously collected, and the audit task template is iteratively optimized. The template content is adjusted according to new audit experience and problems found to ensure that it is always highly consistent with actual audit needs.

[0080] After determining the professional labels of the construction drawing data to be reviewed, the system or auditor will automatically identify these labels and prepare for the subsequent review task formulation. Using the preset mapping relationship or algorithm, the identified professional labels are matched with the corresponding review task templates. This matching process may involve the evaluation and selection of multiple templates to ensure that the final selected template best meets the needs of the current review task. According to the matched review task template, specific review tasks are automatically generated. These tasks will clearly indicate the specific content, steps, standards and expected results that need to be reviewed, providing clear guidance and reference for auditors. Finally, according to the needs of the review task and the expertise of the auditor, the generated review tasks will be assigned to the appropriate auditors for execution. At the same time, the time limit and quality requirements for task completion can be set to ensure the timeliness and effectiveness of the review work.

[0081] See also Figure 3 , Figure 3 The present invention provides Figure 1 In some embodiments of the present invention, extracting audit opinion elements from historical audit opinion data in a tag database corresponding to a professional tag also includes:

[0082] S310. Perform cluster analysis on the historical audit opinion data to determine the basic elements of the historical audit opinion data;

[0083] S320, constructing a knowledge graph based on basic elements, and determining the association relationship between basic elements in the knowledge graph;

[0084] S330. Extract corresponding audit opinion elements from historical audit opinion data based on the association relationship and the construction drawing data to be audited.

[0085] In the above embodiment, first, the historical audit opinion data in the label database is cleaned to remove duplicate, invalid or inconsistently formatted data to ensure the accuracy and efficiency of subsequent analysis. Key features are extracted from the cleaned data. These features may include the specific content of the audit opinion, the regulatory clauses involved, the type of problem, the solution, etc. These features will serve as the basis for cluster analysis. According to the characteristics of the data and the analysis requirements, a suitable clustering algorithm is selected, such as K-means, hierarchical clustering, DBSCAN, etc. The selection of the algorithm should take into account factors such as data distribution, cluster shape, and noise points. The selected clustering algorithm is used to perform cluster analysis on the historical audit opinion data to obtain several clustering results. By evaluating the clustering effect (such as the silhouette coefficient, Calinski-Harabasz index, etc.), the optimal number of clusters and clustering results are determined. According to the clustering results, the basic elements of the historical audit opinion data represented by each cluster are extracted. These basic elements are the key to the subsequent construction of the knowledge graph and the extraction of audit opinion elements.

[0086] Identify specific entities in the basic elements, such as regulatory clauses, problem types, solutions, etc. These entities will serve as nodes in the knowledge graph. Use natural language processing technology or preset rules to extract relationships between entities from historical review opinion data, such as "a certain problem type violates a certain regulatory clause", "a certain solution is suitable for solving a certain problem type", etc. These relationships will serve as edges in the knowledge graph. Integrate the identified entities and extracted relationships together to construct a complete knowledge graph. The graph should clearly show the relationship between the basic elements to facilitate subsequent query and reasoning. Optimize the constructed knowledge graph, including removing redundant information, merging similar entities, adjusting relationship weights, etc., to improve the accuracy and practicality of the graph.

[0087] Match the key information in the construction drawing data to be reviewed with the entities in the knowledge graph to find the historical audit opinion data related to it. Search the knowledge graph for the association paths between the entities that match the construction drawing data to be reviewed. These paths reveal the logical relationship between different basic elements and provide a basis for extracting audit opinion elements. According to the searched association paths and the actual needs of the construction drawing data to be reviewed, extract the corresponding audit opinion elements from the historical audit opinion data. These elements may include possible problems, key points that need attention, and recommended improvement measures. Organize the extracted audit opinion elements into easy-to-understand audit guidance documents or reports to provide auditors with clear audit directions and key points. At the same time, the audit opinion elements can also be associated with the construction drawing data to be reviewed and stored for subsequent query and tracing.

[0088] See also Figure 4 , Figure 4 The present invention provides Figure 3In some embodiments of the present invention, cluster analysis is performed on the historical audit opinion data to determine the basic elements of the historical audit opinion data, and further includes:

[0089] S311. Clean the historical audit opinion data, remove duplicate data and format them into the same data format;

[0090] S312, determining a vocabulary vector sequence according to the frequency of occurrence of each vocabulary in the historical review opinion data;

[0091] S313, clustering the vocabulary vector sequence based on a preset clustering algorithm to obtain element clusters;

[0092] S314. Analyze the element clusters to determine the basic elements represented by each element cluster.

[0093] In the above embodiment, a hash table, database query or a special deduplication algorithm is used to identify and delete duplicate records in the historical audit opinion data to ensure that the data set for subsequent analysis is unique. The cleaned data is uniformly converted into a data format suitable for analysis, such as text format, JSON or CSV, etc. At the same time, the text data is preprocessed, such as removing punctuation, stop word filtering, stemming or word form restoration, etc., to improve the accuracy of subsequent vocabulary vectorization.

[0094] Extract all words from the formatted historical audit opinion data and build a vocabulary. Count the frequency of each word in the data set as an indicator of word importance. Use the word frequency statistics results, combined with natural language processing technologies such as TF-IDF (term frequency-inverse document frequency), Word2Vec, and BERT, to convert the words into vector form. These vectors can capture the semantic relationship and importance differences between words. Convert each historical audit opinion data into a sequence of word vectors as input for subsequent clustering analysis.

[0095] Select appropriate clustering algorithms according to data characteristics and analysis requirements, such as K-means, hierarchical clustering, DBSCAN, spectral clustering, etc. Considering the high dimensionality and possible nonlinear relationship of vocabulary vector sequences, more complex clustering algorithms or combined dimensionality reduction techniques may be required. Set corresponding parameters according to the requirements of the clustering algorithm, such as the number of clusters K, distance measurement method, convergence conditions, etc. The setting of these parameters has an important impact on the clustering results. Use the selected clustering algorithm to cluster the vocabulary vector sequence to obtain several element clusters, each of which represents a set of words with similar semantics or importance.

[0096] Conduct descriptive analysis on each element cluster, and count the co-occurrence of words in the cluster, topic distribution and other characteristics to understand the commonalities and differences of words in the cluster. Based on the descriptive analysis results of the cluster, combined with domain knowledge and audit experience, determine the basic elements represented by each cluster. These basic elements are the key to the subsequent construction of the knowledge graph and the extraction of audit opinion elements. Name the determined basic elements and annotate them with corresponding keywords or phrases to facilitate subsequent representation and query in the knowledge graph.

[0097] See also Figure 5 , Figure 5 The present invention provides Figure 1 In some embodiments of the present invention, the process of generating the audit opinion corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit task also includes:

[0098] S410, automatically filling in a preset audit opinion template based on the audit opinion elements and the audit task to obtain an initial audit opinion;

[0099] S420, using natural language generation technology to adjust the initial audit opinion to obtain a transitional audit opinion;

[0100] S430. Perform sentence detection on the transitional audit opinion, modify the detected sentence problems and obtain the corresponding audit opinion.

[0101] In the above embodiment, one or more audit opinion templates are pre-designed, which are customized according to common audit questions and concerns, and contain fixed sentence structures and replaceable placeholders. According to the requirements of the audit task and the specific content of the audit opinion elements, the elements are matched with the placeholders in the template, which includes determining which elements need to be included in the audit opinion, and their position and order in the opinion. Using automated tools or scripts, the matching audit opinion elements are filled into the preset audit opinion template to generate a preliminary, structured audit opinion.

[0102] Conduct semantic analysis on the initial review opinion to understand its overall meaning and contextual relationship, which helps to identify possible problems of unclear semantics, incoherent logic or inaccurate expression. Based on the results of semantic analysis, use natural language generation technology (such as text rewriting, sentence reorganization, vocabulary replacement, etc.) to adjust the initial review opinion. The purpose of the adjustment is to make the review opinion clearer, more accurate, easier to understand, and in line with professional terminology and expression habits. The review opinion after content adjustment is the transitional review opinion, which retains the basic structure and key points of the initial review opinion, but has been significantly improved in expression and language fluency.

[0103] Using tools such as grammar checking, spelling checking, and semantic consistency checking in natural language processing (NLP) technology, sentence detection is performed on the transitional review opinions, which helps to discover potential grammatical errors, spelling errors, punctuation errors, and semantic inconsistencies or contradictions. Modifications are made one by one for the detected sentence problems, including correcting grammatical errors, correcting spelling errors, adjusting the use of punctuation, and resolving semantic inconsistencies or contradictions. The accuracy and professionalism of the review opinions should be ensured during the modification process. The transitional review opinions after sentence detection and problem modification become the final review opinions, which will serve as formal feedback on the construction drawing data to be reviewed for reference and processing by relevant personnel.

[0104] In order to better implement the audit opinion generation method based on historical data in the embodiment of the present invention, on the basis of the audit opinion generation method based on historical data, correspondingly, please refer to Figure 6 , Figure 6 This is a schematic diagram of the structure of an embodiment of a device for generating audit opinions based on historical data provided by the present invention. The embodiment of the present invention provides a device for generating audit opinions based on historical data 600, including:

[0105] A classification storage module 610, which is configured to classify the historical audit opinion data using a well-trained audit opinion classification model and store it in a corresponding label database;

[0106] A task formulation module 620 is configured to determine the professional tags that need to be reviewed from the construction drawing data to be reviewed, and formulate the review tasks according to the professional tags;

[0107] An element extraction module 630 is configured to extract audit opinion elements from historical audit opinion data in a tag database corresponding to a professional tag;

[0108] The opinion generating module 640 is configured to generate an audit opinion corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit task.

[0109] It should be noted here that: the device 600 provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above method embodiments, which will not be repeated here.

[0110] See also Figure 7 , Figure 7A schematic diagram of the structure of a device for generating audit opinions based on historical data provided by an embodiment of the present invention. Based on the above-mentioned method for generating audit opinions based on historical data, the present invention also provides a device for generating audit opinions based on historical data. The device for generating audit opinions based on historical data can be a computing device such as a mobile terminal, a desktop computer, a notebook, a PDA, and a server. The device for generating audit opinions based on historical data 700 includes a processor 710, a memory 720, and a display 730. Figure 7 Only some components of the synchronous tracking flying welding equipment for real-time measurement of battery height are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0111] In some embodiments, the memory 720 may be an internal storage unit of the audit opinion generating device 700 based on historical data, such as a hard disk or memory of the audit opinion generating device 700 based on historical data. In other embodiments, the memory 720 may also be an external storage device of the audit opinion generating device 700 based on historical data, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the audit opinion generating device 700 based on historical data. Further, the memory 720 may also include both the internal storage unit of the audit opinion generating device 700 based on historical data and an external storage device. The memory 720 is used to store application software and various types of data installed in the audit opinion generating device 700 based on historical data, such as the program code of the audit opinion generating device 700 based on historical data. The memory 720 may also be used to temporarily store data that has been output or is to be output. In one embodiment, the memory 720 stores an audit opinion generation program 740 based on historical data, and the audit opinion generation program 740 based on historical data can be executed by the processor 710, thereby realizing the audit opinion generation method based on historical data of each embodiment of the present application.

[0112] In some embodiments, the processor 710 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 720, such as executing a method for generating audit opinions based on historical data.

[0113] In some embodiments, the display 730 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 730 is used to display information on the audit opinion generating device based on historical data 700 and to display a visual user interface. The components 710-730 of the audit opinion generating device based on historical data 700 communicate with each other via a system bus.

[0114] In one embodiment, when the processor 710 executes the audit opinion generation program 740 based on historical data in the memory 720 , the steps in the above method for generating an audit opinion based on historical data are implemented.

[0115] This embodiment further provides a computer-readable storage medium on which a path planning program is stored. When the path planning program is executed by a processor, the following steps are implemented:

[0116] Use the well-trained audit opinion classification model to classify historical audit opinion data and store them in the corresponding label database;

[0117] Determine the professional labels that need to be reviewed from the construction drawing data to be reviewed, and formulate review tasks based on the professional labels;

[0118] Extracting audit opinion elements from historical audit opinion data in the label database corresponding to the professional label;

[0119] Generate audit opinions corresponding to the construction drawing data to be audited based on the audit opinion elements and audit tasks.

[0120] In summary, the present invention provides a method for generating audit opinions based on historical data. By classifying and storing historical audit opinions, the system can quickly find relevant historical audit opinion data according to the professional labels of the construction drawings to be audited. This method significantly reduces the time for manually reviewing a large number of historical documents, making the audit process more efficient. Using a well-trained audit opinion classification model, it can ensure that historical audit opinions are accurately classified and stored. At the same time, the audit opinions generated based on these verified historical data are more likely to cover common audit points and potential problems, thereby improving the accuracy and comprehensiveness of the audit. Through data analysis, the system can discover common problems and high-frequency problems in the audit process, and provide data support for the formulation of more scientific and reasonable audit strategies and standards. At the same time, the prediction model based on big data can also help predict potential risks. It helps to realize the automation and intelligence of the audit process, reduce manual intervention links, and reduce the probability of human errors. At the same time, through real-time monitoring and feedback mechanisms, problems in the audit process can be discovered and corrected in a timely manner, and the audit process can be continuously optimized.

[0121] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0122] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0123] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0124] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0125] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0126] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0127] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program code.

[0128] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0129] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0130] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0131] It will be easily understood by those skilled in the art that 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 and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating audit opinions based on historical data, characterized in that: include: Use the well-trained audit opinion classification model to classify historical audit opinion data and store them in the corresponding label database; Determine the professional tags that need to be reviewed from the construction drawing data to be reviewed, and formulate the review tasks according to the professional tags; Extracting audit opinion elements from the historical audit opinion data in the tag database corresponding to the professional tag; An audit opinion corresponding to the construction drawing data to be audited is generated based on the audit opinion elements and the audit task.

2. The method for generating audit opinions based on historical data according to claim 1, characterized in that: The method of using a well-trained audit opinion classification model to classify historical audit opinion data and store them in a corresponding label database also includes: Obtain a fully trained audit opinion classification model; Inputting the historical audit opinion data into the fully trained audit opinion classification model, and outputting professional labels corresponding to the historical audit opinion data; A label database is established according to the professional label, and the historical review opinion data corresponding to the professional label is stored in the label database.

3. The method for generating audit opinions based on historical data according to claim 1, characterized in that: The construction drawing data to be reviewed includes image data and text data; the step of determining the professional tags to be reviewed from the construction drawing data to be reviewed also includes: Performing image recognition on the image data using a graph neural network model to extract key factors from the image data; The text data is semantically analyzed based on the key factors to determine professional labels for the construction drawing data to be reviewed.

4. The method for generating audit opinions based on historical data according to claim 3, characterized in that: The formulating of the audit task according to the professional label also includes: Formulate audit task templates based on the historical audit opinion data; The professional tag is matched with the audit task template to obtain an audit task.

5. The method for generating audit opinions based on historical data according to claim 1, characterized in that: The extracting of audit opinion elements from the historical audit opinion data in the tag database corresponding to the professional tag also includes: Performing cluster analysis on the historical audit opinion data to determine the basic elements of the historical audit opinion data; Construct a knowledge graph based on the basic elements, and determine the association relationship between the basic elements in the knowledge graph; Corresponding audit opinion elements are extracted from the historical audit opinion data according to the association relationship and the construction drawing data to be audited.

6. The method for generating audit opinions based on historical data according to claim 5, characterized in that: The cluster analysis of the historical audit opinion data to determine the basic elements of the historical audit opinion data also includes: Clean the historical audit opinion data, remove duplicate data and format them into the same data format; Determine a vocabulary vector sequence according to the frequency of occurrence of each vocabulary in the historical review opinion data; Clustering the vocabulary vector sequence based on a preset clustering algorithm to obtain element clusters; The element clusters are analyzed to determine the basic elements represented by each element cluster.

7. The method for generating audit opinions based on historical data according to claim 1, characterized in that: The step of generating the audit opinion corresponding to the to-be-audited construction drawing data based on the audit opinion elements and the audit task also includes: Automatically filling in a preset audit opinion template based on the audit opinion elements and the audit task to obtain an initial audit opinion; Using natural language generation technology to adjust the initial audit opinion to obtain a transitional audit opinion; The transitional audit opinion is subjected to statement detection, and the detected statement problems are modified to obtain the corresponding audit opinion.

8. A device for generating audit opinions based on historical data, characterized in that: include: A classification storage module, which is configured to classify the historical audit opinion data using a well-trained audit opinion classification model and store it in a corresponding label database; A task formulation module, which is configured to determine the professional tags that need to be reviewed from the construction drawing data to be reviewed, and formulate the review task according to the professional tags; An element extraction module, configured to extract audit opinion elements from the historical audit opinion data in the tag database corresponding to the professional tag; An opinion generating module is configured to generate an audit opinion corresponding to the construction drawing data to be audited based on the audit opinion elements and the audit task.

9. A device for generating audit opinions based on historical data, characterized in that: It includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method for generating audit opinions based on historical data as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: It stores a computer program executable by an access authentication device. When the computer program runs on the access authentication device, the access authentication device executes the steps of the method for generating audit opinions based on historical data as described in any one of claims 1 to 7.

Images