Project information analysis and duplicate checking method

Abstract

The invention discloses a project information analysis and duplicate checking method, which comprises the following steps of: 1, importing project information which comprises a bidding file and other project files, selecting different processing modes to analyze the project information according to the type of the project information, and converting the project information into a standard form; 2, performing filling inspection after conversion into a standard form, performing semantic recognition after the filling inspection is passed, obtaining a filling content analysis result, and generating prompt information if the filling inspection is not passed, and prompting to perform original project information inspection; 3, when the filled analysis result is not abnormal, project information analysis is carried out, and a project analysis result is obtained; 4, analyzing the project information; 5, after project information duplicate checking is completed, duplicate checking information is generated and sent to a preset receiving terminal; according to the invention, project information analysis and duplicate checking are carried out more accurately.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and particularly to a method for project information analysis and duplicate checking. Background Art

[0002] With the booming development of the intelligent construction field, various projects show the characteristics of large scale, high complexity, and numerous participants. From the construction of large-scale smart city infrastructure to the construction of intelligent residential communities and smart factories, a vast amount of project information needs to be efficiently processed and precisely controlled.

[0003] The formats of project information from different sources are diverse. There are both traditional paper documents and electronic files, such as tender documents, various forms, handwritten annotations on construction drawings, and paper archives of various approval forms. There are also electronic documents, such as progress reports generated by project management software, electronic drawings delivered by design teams, etc.

[0004] In order to ensure the quality of project information, project information analysis and duplicate checking are required. Therefore, a method for project information analysis and duplicate checking is proposed. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method for project information analysis and duplicate checking, including the following steps:

[0006] Step 1: Import project information, where the project information includes bidding documents and other project files. Different processing methods are selected according to the type of project information to analyze the project information and convert it into a standard form;

[0007] Step 2: After converting to the standard form, perform a filling check. After the filling check passes, perform semantic recognition to obtain the analysis result of the filled content. If the filling check fails, generate a prompt message to prompt for the inspection of the original project information;

[0008] Step 3: When there is no abnormality in the filling analysis result, perform project information analysis to obtain the project analysis result;

[0009] Step 4: Then, analyze the project information again, extract keywords, and perform duplicate checking on the project information after extraction;

[0010] Step 5: After the project information duplicate checking is completed, generate duplicate checking information and send it to a preset receiving terminal.

[0011] Furthermore, the specific process in Step 1 is as follows:

[0012] The types of other project files include paper handwritten project files, paper printed project files, and electronic document version project files;

[0013] When the type of other project documents is paper handwritten project documents:

[0014] First, based on the optical character recognition method of deep learning, perform preliminary recognition to convert handwritten text into preliminary recognition results;

[0015] When there are suspected errors or unclear characters in the preliminary recognition results, use the optical character recognition method based on semantic understanding for secondary recognition verification. The system automatically captures the text content before and after the character, combines the industry domain knowledge graph and semantic analysis algorithm for correction, and marks the corrected content to obtain the secondary recognition results. If there are no suspected errors or unclear characters, directly proceed to the next step;

[0016] Finally, conduct manual inspection through the manual proofreading module. If the manual verification passes, export the final recognition results.

[0017] When the type of other project documents is paper printed project documents:

[0018] After scanning the paper document into a digital image, perform hierarchical optical character recognition operations;

[0019] In the first layer, use a high-precision general optical character recognition engine to quickly extract text content and identify the main body of the text and basic format information;

[0020] In the second layer, enable the intelligent optical character recognition module to adopt different recognition strategies for different elements such as titles, paragraphs, tables, and charts;

[0021] Finally, conduct two-way format verification. On the one hand, from the perspective of visual presentation, check whether the layout of each element in the digital image conforms to the industry's conventional format specifications;

[0022] On the other hand, from the perspective of the logic of the text content, compare whether the information between different elements is consistent;

[0023] Finally, export the recognition results that pass the two-way format verification;

[0024] When the type of other project documents is electronic document version project documents, directly convert them into standard forms;

[0025] When the project information is bidding documents, the bidding documents include paper bidding documents and electronic bidding documents. For paper bidding documents, perform the same processing as paper printed project documents. For electronic bidding documents, directly convert them into standard forms.

[0026] Furthermore, when the project information is other project documents, the types of other project documents include paper handwritten project documents and paper printed project documents. Before the recognition of paper handwritten project documents, flatness detection needs to be carried out. After the flatness detection passes, the next step of recognition is allowed.

[0027] Furthermore, the specific process of flatness detection is as follows:

[0028] First, image acquisition is carried out, that is, using an image acquisition device, the image information of the paper handwritten project document is acquired at 45 degrees downward from the left front, 45 degrees downward from the right front, 45 degrees downward from directly above, and 45 degrees downward from directly below, obtaining the first image information, the second image information, the third image information, and the fourth image information;

[0029] After that, the left front edge center point T1, the right front edge center point T2, the directly above edge center point T3, and the directly below edge center point T4 are extracted from the first image information;

[0030] After that, taking the plane where the paper handwritten project document is placed as the reference plane, the distance q1 between T1 and the reference plane, the distance m1 between T2 and the reference plane, the distance f1 between T3 and the reference plane, and the distance e1 between T4 and the reference plane are measured;

[0031] The q1, m1, f1, and e1 are combined to obtain the first evaluation parameter A1(q1, m1, f1, e1);

[0032] Then, the second image information, the third image information, and the fourth image information are processed in the same way as the first image information to obtain the second evaluation parameter A2(q2, m2, f2, e2), the third evaluation parameter A3(q3, m3, f3, e3), and the fourth evaluation parameter A4(q4, m4, f4, e4);

[0033] After that, the difference Aa1 between the first evaluation parameter and the second evaluation parameter, and the difference Aa2 between the third evaluation parameter and the fourth evaluation parameter are calculated;

[0034] The difference Aa3 between the first evaluation parameter and the third evaluation parameter, the difference Aa4 between the first evaluation parameter and the fourth evaluation parameter, the difference Aa5 between the second evaluation parameter and the fourth evaluation parameter, and the difference Aa6 between the second evaluation parameter and the third evaluation parameter;

[0035] When at least four or more of Aa1, Aa2, Aa3, Aa4, Aa5, and Aa6 exceed the preset range, it means that the flatness detection fails.

[0036] Furthermore, the specific process of manual inspection through the manual proofreading module is as follows:

[0037] When there is a second verification result, at least two people need to be arranged to confirm the suspected incorrect or unclear characters at least twice. When the two confirmation results are the same, export is allowed.

[0038] When the two confirmation results are different, re-verification is carried out.

[0039] Furthermore, the specific process of the above-mentioned filling check is as follows:

[0040] First, perform text area recognition, collect the image information after filling, locate the position of the filling box in the image, and obtain its upper left coordinate (x1, y1) and lower right coordinate (x2, y2).

[0041] Then, identify the range of the text within the filling box, and also obtain the upper left coordinate (x3, y3) and lower right coordinate (x4, y4) of the text area.

[0042] After that, calculate the center coordinates of the text area, the abscissa x 中 =(x3 + x4) / 2, and the ordinate y 中 =(y3 + y4) / 2;

[0043] Next, calculate the center coordinates of the filling box, the abscissa x 框中 =(x1 + x2) / 2, and the ordinate y 框中 =(y1 + y2) / 2;

[0044] Calculate the abscissa deviation value Δx = |x 中 -x 框中 | and the ordinate deviation value Δy = |y 中 -y 框中 |;

[0045] When both the abscissa deviation value Δx and the ordinate deviation value Δy are less than the set threshold, it is considered that the text position is centered, which means the filling check passes.

[0046] Furthermore, the project analysis results include project anomalies and no project anomalies;

[0047] The process of obtaining the project analysis results is as follows:

[0048] Extract the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem from the project information;

[0049] Process the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem to obtain the project progress indicators;

[0050] Then extract the proportion of batches with unqualified material quality inspections, the rework rate, and the average time between failures of the intelligent system from the project information;

[0051] Process the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system to obtain quality control indicators;

[0052] Then extract the cost-benefit ratio, budget change frequency and amplitude, and cost deviation rate from the project information;

[0053] Process the cost-benefit ratio, budget change frequency and amplitude, and cost deviation rate to obtain cost budget indicators;

[0054] Process the project progress indicators, quality control indicators, and cost budget indicators to obtain comprehensive evaluation indicators. When the comprehensive evaluation indicator is less than the preset value, it indicates that the project is abnormal; otherwise, it indicates that the project is normal.

[0055] Furthermore, the process of obtaining project progress indicators is as follows: First, establish a mapping set of engineering progress score values for the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem. After obtaining the specific score values of the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem from the mapping set of engineering progress score values, assign different weights to the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem, and then calculate the sum of the weighted score values to obtain the project progress indicators;

[0056] The process of obtaining quality control indicators is as follows:

[0057] First, establish a mapping set of quality control score values for the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system. After obtaining the specific score values of the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system from the mapping set of quality control score values, assign different weights to the proportion of unqualified batches in building material quality inspection, the rework rate, and the mean time between failures of the intelligent system, and then calculate the sum of the weighted score values to obtain the quality control indicators;

[0058] The process of obtaining cost budget indicators is as follows:

[0059] First, establish a mapping set of cost budget score values for the cost-benefit ratio, budget change frequency and amplitude, and cost deviation rate. After obtaining the specific score values of the cost-benefit ratio, budget change frequency and amplitude, and cost deviation rate from the mapping set of cost budget score values, assign different weights to the cost-benefit ratio, budget change frequency and amplitude, and cost deviation rate, and then calculate the sum of the weighted score values to obtain the cost budget indicators;

[0060] The process of obtaining comprehensive evaluation indicators is as follows:

[0061] Mark the project progress indicators as G1, mark the quality control indicators as G2, and mark the cost budget indicators as G3;

[0062] Assign weights W1 to G1, W2 to G2, and W3 to G3, where W1 + W2 + W3 = 1 and W3 > W2 = W1;

[0063] By using the formula G1 * W1 + G2 * W2 + G3 * W3 = Gg, the comprehensive evaluation index Gg is obtained.

[0064] Furthermore, the duplicate checking process in step four is as follows:

[0065] First, perform keyword extraction: Clean the project information text to remove the noise data therein, and obtain the preprocessed text;

[0066] After that, use a statistics-based method: Apply the term frequency-inverse document frequency (TF-IDF) algorithm to calculate the importance scores of each word in the project information text;

[0067] Then, use a rule-based method combined with the domain knowledge and business rules to which the project information belongs to formulate keyword extraction rules;

[0068] Next, perform part-of-speech tagging and screening: Perform part-of-speech tagging on the preprocessed text to identify words of different parts of speech such as nouns, verbs, and adjectives;

[0069] Keyword screening and sorting: After obtaining the candidate keywords according to the above process, screen the keywords according to the set threshold (such as the TF-IDF score threshold) to remove the words with lower scores;

[0070] At the same time, sort the screened keywords according to their importance scores, and select several keywords with higher rankings as the final project information keywords;

[0071] Perform duplicate checking on the project information and establish a project information database: Store the existing project information in the database, and each project information has a unique identifier;

[0072] Establish an index for the keywords extracted from each project information in the database;

[0073] For the project information to be checked for duplicates, extract keywords according to the set keyword extraction rules, and match the keywords of the project information to be checked for duplicates with the keywords of the existing project information in the database;

[0074] Judge whether there is duplicate project information according to the matching results. If a project information highly matching the keywords of the project information to be checked for duplicates is found in the database, it is considered that there is a duplicate; otherwise, it is considered that there is no duplicate for this project information;

[0075] After all keyword matches are completed, generate duplicate checking information.

[0076] Furthermore, during the duplicate check in Step 4, when the imported project information is a bidding document, the specific duplicate check process is as follows:

[0077] Import the bidding document into the preset standard form library for content reading:

[0078] Then, according to the data entry specifications of the preset standard library, store the read content line by line, paragraph by paragraph, or corresponding to specific fields into the standard library, and record the basic information of the file, including the file name, upload time, and file source;

[0079] After that, perform a search for the compliant part. According to the preset standard rules, the standard rules exist in the metadata area of the standard library in the forms of text descriptions, templates, and keyword sets;

[0080] After that, through text matching and pattern recognition technologies, screen out the paragraphs, clauses, and data items that match the standard rules in the content of the imported bidding document, mark them as compliant, and generate a search report to record the specific locations and content summaries of the compliance;

[0081] Then, extract the non-compliant part. Compare the content that has been retrieved as compliant, and use the difference comparison algorithm to find the different parts in the bidding document. The different parts include text expressions, numerical ranges, and format layouts;

[0082] Extract the different parts, that is, the difference parts, separately and organize them into an independent data set;

[0083] Input the extracted difference parts into a pre-trained deep learning model. The deep learning model analyzes the difference parts from semantic understanding, logical structure, and industry practices, judges whether they meet the requirements of industry general norms, and outputs the analysis results. The analysis results include compliant and non-compliant;

[0084] For the difference parts with the analysis result of non-compliant, the system automatically extracts the difference features;

[0085] Record the difference features in a structured form and store them in the violation feature library of the preset standard library;

[0086] When there are parts of the content that the deep learning model cannot determine, perform manual determination on this part of the content. When the manual determination is non-compliant, extract the features of this part again and import them into the violation feature library of the preset standard library.

[0087] The beneficial effects of the present invention are reflected in:

[0088] It can process different types of project information such as paper handwritten, paper printed, and electronic document versions, meeting the diverse information source requirements in intelligent construction, ensuring that various forms of project information can enter the analysis process, and improving the generality and applicability of the method.

[0089] For paper handwritten project documents, a combination of deep learning-based optical character recognition, secondary recognition verification based on semantic understanding, and manual proofreading is adopted, effectively overcoming the difficulties of handwritten text recognition and greatly improving the recognition accuracy; for paper printed project documents, hierarchical optical character recognition and two-way format verification ensure the quality of the recognition results; electronic document version project documents are directly converted into a standard form, which is also convenient for subsequent unified processing.

[0090] Before recognizing paper handwritten project documents, the flatness detection is carried out. Through multi-angle acquisition of image information and complex calculation and evaluation, the problem of uneven paper can be discovered in advance, avoiding the influence of paper deformation on the recognition results and laying a foundation for subsequent accurate information extraction and analysis.

[0091] By accurately recognizing the text area and calculating the coordinates, it is judged whether the filled content is centered, and the problem of non-standard filling can be discovered in time and prompt information is generated, which helps to improve the input quality of project information and ensure the standardization and consistency of data.

[0092] Multiple key indicators are extracted from three dimensions of project progress, quality control, and cost budget to comprehensively evaluate the project, which can comprehensively and deeply reflect the actual situation of intelligent construction projects and provide rich and valuable decision-making basis for project managers.

[0093] Quantitative evaluation improves scientificity: A score value mapping set is established in advance for each dimension index. By assigning different weights and calculating the total score value, the project evaluation index is quantified, making the evaluation results more objective and scientific, reducing the interference of subjective factors, and improving the accuracy and credibility of the evaluation.

[0094] Adopting a keyword extraction strategy combining multiple methods can accurately and comprehensively extract representative keywords from the project information text. These keywords can highly summarize the core content of the project and provide a reliable basis for duplicate checking work.

[0095] By establishing a project information database and indexing the keywords, the keywords of the project information to be checked for duplicates are matched with the existing information in the database, which can quickly and accurately judge whether the project information is repeated, helping to avoid duplicate construction, improve the resource utilization efficiency, and ensure the uniqueness and innovation of intelligent construction projects. Description of the Drawings

[0096] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0097] Figure 1 is the overall flowchart of the present invention;

[0098] Figure 2 is the schematic diagram for collecting the flatness detection image of the present invention. Specific Embodiments

[0099] The following will describe in detail the embodiments of the technical solutions of the present invention in conjunction with the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention, so they are only examples and cannot be used to limit the protection scope of the present invention.

[0100] It should be noted that unless otherwise specified, the technical terms or scientific terms used in this application should have the ordinary meanings understood by those skilled in the art to which the present invention belongs.

[0101] As Figures 1 to 2 shown, a method for project information analysis and duplicate checking includes the following steps:

[0102] Step 1: Import project information, where the project information includes bidding documents and other project documents. Different processing methods are selected according to the types of project information to analyze the project information and convert it into a standard form;

[0103] Step 2: After converting to the standard form, perform a filling check. After the filling check passes, perform semantic recognition to obtain the analysis result of the filled content. If the filling check fails, a prompt message is generated to prompt for the inspection of the original project information;

[0104] Step 3: When there is no abnormality in the filling analysis result, perform project information analysis to obtain the project analysis result;

[0105] Step 4: Then analyze the project information again, extract keywords, and perform duplicate checking of the project information after extraction;

[0106] Step 5: After the duplicate checking of the project information is completed, generate duplicate checking information and send it to the preset receiving terminal.

[0107] Furthermore, the specific process in Step 1 is as follows:

[0108] The types of other project documents include paper handwritten project documents, paper printed project documents, and electronic document version project documents;

[0109] When the type of other project documents is paper handwritten project documents:

[0110] First, based on the optical character recognition method of deep learning, perform preliminary recognition to convert the handwritten text into a preliminary recognition result;

[0111] When there are suspected errors or unclear characters in the preliminary recognition result, use the optical character recognition method based on semantic understanding for secondary recognition verification. The system automatically grabs the text content before and after the character, combines the industry domain knowledge graph and semantic analysis algorithm for correction, and marks the corrected content to obtain the secondary recognition result. If there are no suspected errors or unclear characters, directly proceed to the next step;

[0112] Finally, conduct manual inspection through the manual proofreading module. If the manual verification passes, export the final recognition result.

[0113] When the type of other project documents is paper printed project documents:

[0114] After scanning the paper document into a digital image, perform a hierarchical optical character recognition operation;

[0115] In the first layer, use a high-precision general optical character recognition engine to quickly extract the text content and identify the text body and basic format information;

[0116] In the second layer, enable an intelligent optical character recognition module focused on complex layout parsing, and adopt different recognition strategies for different elements such as titles, paragraphs, tables, and charts;

[0117] When recognizing the title, combine the industry knowledge graph and the title hierarchy relationship model to judge its importance level in the entire project document architecture. For example, the first-level title usually covers the core theme of the project, and the second-level title refines the key sections under the core theme, etc.; for paragraphs, in addition to analyzing the hierarchy based on paragraph indentation and line spacing changes, also use text classification algorithms to classify the paragraph content into different categories such as background introduction, technical solutions, implementation plans, etc.; for tables, build a table semantic understanding model to not only ensure the accurate transfer of row and column data to the standard format table filling area, but also interpret the logical relationship behind the table data. For example, in the project resource allocation table, analyze whether the proportion of resources allocated to different departments is reasonable and whether it conforms to the overall project plan.

[0118] Finally, perform a two-way format verification. On the one hand, from the perspective of visual presentation, check whether the layout of each element in the digital image conforms to the industry's conventional format specifications, such as whether the table lines are neat and whether the axis labels of the chart are clear, etc.;

[0119] On the other hand, from the perspective of the logical relationship of the text content, check whether the information between different elements is consistent. For example, check whether the project scope mentioned in the title matches the scope reflected in the paragraph description and table data. Ensure the accuracy and integrity of the information through two-way verification. After passing the comparison, fill the accurately parsed project information into the corresponding filling boxes according to the standard format requirements;

[0120] Finally, export the recognition results that have passed the two-way format verification;

[0121] When the type of other project documents is electronic document project files, directly convert them into the standard form;

[0122] When the project information is bidding documents, the bidding documents include paper bidding documents and electronic bidding documents. For paper bidding documents, perform the same processing as for paper printed project documents. For electronic bidding documents, directly convert them into the standard form.

[0123] When the project information is other project documents, the types of other project documents include paper handwritten project documents and paper printed project documents. Before identifying paper handwritten project documents, it is necessary to perform flatness detection. After passing the flatness detection, the next step of identification is allowed.

[0124] The specific process of performing flatness detection is as follows:

[0125] First, perform image acquisition, that is, use an image acquisition device to acquire the image information of the paper handwritten project document at 45 degrees downward from the left front, 45 degrees downward from the right front, 45 degrees downward from the directly above, and 45 degrees downward from the directly below, and obtain the first image information, the second image information, the third image information, and the fourth image information;

[0126] After that, extract the center point T1 of the left edge, the center point T2 of the right edge, the center point T3 of the upper edge, and the center point T4 of the lower edge from the first image information;

[0127] After that, take the plane on which the paper handwritten project document is placed as the reference plane, and measure the distance q1 between T1 and the reference plane, the distance m1 between T2 and the reference plane, the distance f1 between T3 and the reference plane, and the distance e1 between T4 and the reference plane;

[0128] Combine q1, m1, f1, and e1 to obtain the first evaluation parameter A1(q1, m1, f1, e1);

[0129] Then perform the same processing on the second image information, the third image information, and the fourth image information as on the first image information to obtain the second evaluation parameter A2(q2, m2, f2, e2), the third evaluation parameter A3(q3, m3, f3, e3), and the fourth evaluation parameter A4(q4, m4, f4, e4);

[0130] After that, calculate the difference Aa1 between the first evaluation parameter and the second evaluation parameter, and the difference Aa2 between the third evaluation parameter and the fourth evaluation parameter;

[0131] The difference Aa3 between the first evaluation parameter and the third evaluation parameter, the difference Aa4 between the first evaluation parameter and the fourth evaluation parameter, the difference Aa5 between the second evaluation parameter and the fourth evaluation parameter, and the difference Aa6 between the second evaluation parameter and the third evaluation parameter;

[0132] When at least four or more of Aa1, Aa2, Aa3, Aa4, Aa5, and Aa6 exceed the preset range, it means that the flatness detection fails.

[0133] Furthermore, the specific process of manual inspection through the manual verification module is as follows:

[0134] When there is a second verification result, at least two people need to be arranged to confirm the suspected errors or unclear characters at least twice. When the two confirmation results are the same, export is allowed;

[0135] When the two confirmation results are different, re-verification is performed.

[0136] Furthermore, the specific process of the above-mentioned filling check is as follows:

[0137] First, perform text area recognition, collect the image information after filling, locate the position of the filling box in the image, and obtain its upper left coordinate (x1, y1) and lower right coordinate (x2, y2);

[0138] Then, recognize the range of the text in the filling box, and also obtain the upper left coordinate (x3, y3) and lower right coordinate (x4, y4) of the text area;

[0139] After that, calculate the center coordinate of the text area, the abscissa x 中 =(x3 + x4) / 2, and the ordinate y 中 =(y3 + y4) / 2;

[0140] Next, calculate the center coordinate of the filling box, the abscissa x 框中 =(x1 + x2) / 2, and the ordinate y 框中 =(y1 + y2) / 2;

[0141] Calculate the abscissa deviation value Δx = |x 中 - x 框中 | and the ordinate deviation value Δy = |y 中 - y 框中 |;

[0142] When both the horizontal coordinate deviation value Δx and the vertical coordinate deviation value Δy are less than the set threshold, it is considered that the text position is centered, indicating that the filling check has passed.

[0143] The project analysis results include project anomalies and no project anomalies;

[0144] The process of obtaining the project analysis results is as follows:

[0145] Extract the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem from the project information;

[0146] Process the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem to obtain the project progress indicators;

[0147] Then extract the proportion of batches with unqualified material quality inspections, the rework rate, and the average time between failures of the intelligent system from the project information;

[0148] Process the proportion of batches with unqualified material quality inspections, the rework rate, and the average time between failures of the intelligent system to obtain the quality control indicators;

[0149] Then extract the cost-benefit ratio, the frequency and range of budget changes, and the cost deviation rate from the project information;

[0150] Process the cost-benefit ratio, the frequency and range of budget changes, and the cost deviation rate to obtain the cost budget indicators;

[0151] Process the project progress indicators, the quality control indicators, and the cost budget indicators to obtain the comprehensive evaluation indicators. When the comprehensive evaluation indicator is less than the preset value, it indicates that the project is abnormal; otherwise, it indicates that the project is normal.

[0152] The process of obtaining the project progress indicators is as follows: First, establish a mapping set of project progress score values for the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem. After obtaining the specific score values of the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem from the mapping set of project progress score values, assign different weights to the on-time completion rate of key nodes, the average daily construction volume, and the progress coordination of each subsystem, and then calculate the sum of the weighted scores to obtain the project progress indicators;

[0153] The process of obtaining the quality control indicators is as follows:

[0154] A quality control score value mapping set for the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system has been established in advance. After obtaining the specific values of the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system from the control score value mapping set, different weights are assigned to the proportion of unqualified batches in material quality inspection, the rework rate, and the mean time between failures of the intelligent system, and then the sum of the weighted scores is calculated to obtain the quality control index;

[0155] The process of obtaining the cost budget index is as follows:

[0156] A cost budget score value mapping set for the cost-benefit ratio, the frequency and amplitude of budget changes, and the cost deviation rate has been established in advance. After obtaining the specific values of the cost-benefit ratio, the frequency and amplitude of budget changes, and the cost deviation rate from the cost budget score value mapping set, different weights are assigned to the cost-benefit ratio, the frequency and amplitude of budget changes, and the cost deviation rate, and then the sum of the weighted scores is calculated to obtain the cost budget index;

[0157] The process of obtaining the comprehensive evaluation index is as follows:

[0158] Mark the project progress index as G1, mark the quality control index as G2, and mark the cost budget index as G3;

[0159] Assign weights W1 to G1, W2 to G2, and W3 to G3, where W1 + W2 + W3 = 1 and W3 > W2 = W1;

[0160] Through the formula G1 * W1 + G2 * W2 + G3 * W3 = Gg, the comprehensive evaluation index Gg is obtained.

[0161] Project progress index and scoring:

[0162] On-time completion rate of key nodes: Statistics are made on whether important nodes such as the completion of infrastructure construction, the topping out of the main structure, and the initial commissioning of the intelligent system are completed on time, and the number of nodes completed on time is divided by the total number of key nodes.

[0163] Average daily construction volume: It is calculated by dividing the workload completed in each construction stage (such as building area, number of installed equipment, etc.) by the actual construction days, reflecting the construction efficiency.

[0164] Progress coordination of each subsystem: Measure the matching degree of the progress among subsystems such as building main body construction, mechanical and electrical installation, and intelligent integration. The smaller the difference, the higher the coordination, which can be quantified by calculating the standard deviation of the progress differences of each subsystem.

[0165] Scoring process:

[0166] On-time completion rate of key nodes: 20 points will be given if the completion rate reaches 95% or above; 15 - 20 points will be given if it is between 85% - 94%; 10 - 15 points will be given if it is between 75% - 84%; 0 - 10 points will be given if it is below 75%.

[0167] Average daily construction volume: Compared with the historical average daily construction volume of similar projects, 15 points will be given if it reaches or exceeds the average by 10%; 10 - 15 points will be given if it is within the range of ±10% of the average; 0 - 10 points will be given if it is below the average by -10%.

[0168] Progress coordination of each subsystem: 15 points will be given if the standard deviation is less than or equal to 5%; 10 - 15 points will be given if the standard deviation is between 5% - 10%; 0 - 10 points will be given if the standard deviation is greater than 10%.

[0169] Quality control indicators and scoring:

[0170] Proportion of batches with unqualified material quality inspection: Calculate the proportion of batches with unqualified inspection among various building materials, intelligent components, etc. in the total number of inspected batches.

[0171] Rework rate: Calculate the ratio of the project volume reworked due to quality problems to the total project volume, covering rework caused by unqualified construction technology, equipment commissioning failures, etc.

[0172] Intelligent system operation stability indicator - Mean Time Between Failures (MTBF): Record the average interval duration between two adjacent failures of the intelligent system during a certain operation period. The longer it is, the higher the stability.

[0173] Scoring process:

[0174] Proportion of batches with unqualified material quality inspection: 15 points will be given if the proportion is 0; 1 point will be deducted for every 1% increase until the deduction is completed.

[0175] Rework rate: 15 points will be given if the rework rate is below 3%; 10 - 15 points will be given if the rework rate is between 3% - 5%; 0 - 10 points will be given if the rework rate is above 5%.

[0176] Mean Time Between Failures (MTBF): 20 points will be given if it reaches or exceeds 1.2 times the average MTBF value of similar systems; 10 - 19 points will be given if it is within the range of 1 - 1.2 times the average; 0 - 9 points will be given if it is below the average.

[0177] Cost budget indicators and scoring:

[0178] Cost - benefit ratio: Obtained by dividing the expected project revenue by the total cost, which measures the input - output efficiency.

[0179] Budget change frequency and amplitude: Count the number of budget changes and the percentage of the changed amount in the original budget, which reflects the accuracy of budget planning and control level.

[0180] Cost deviation rate: Calculate the ratio of the difference between the actual cost and the budgeted cost to the budgeted cost. A positive value indicates overspending, and a negative value indicates savings.

[0181] Scoring process:

[0182] Cost-benefit ratio: 20 points are given if it is 10% higher than the industry average; 2 points are deducted for every 5% lower.

[0183] Budget change frequency and amplitude: 15 points are given if the frequency is less than 2 times per year and the amplitude is less than 5%; 10 - 15 points are given if the frequency is 2 - 3 times or the amplitude is 5% - 10%; 0 - 10 points are given if the frequency exceeds 3 times or the amplitude exceeds 10%.

[0184] Cost deviation rate: 15 points are given if the deviation rate is within ±3%; 10 - 15 points are given if the deviation rate is between -3% and -10% or 3% and 10%; 0 - 10 points are given if the deviation rate exceeds the above range.

[0185] The duplicate checking process in Step 4 is as follows:

[0186] First, perform keyword extraction: Clean the project information text, remove the noise data in it, and obtain the preprocessed text.

[0187] Such as HTML tags, special characters, stop words (such as words with no practical meaning in the text like "de", "le", "shi", etc.), convert the text into a unified format, such as lowercase letter form, for subsequent processing.

[0188] After that, use the term frequency-inverse document frequency (TF-IDF) algorithm to calculate the importance score of each word in the project information text.

[0189] Term frequency (TF) represents the frequency of a word in the text, and inverse document frequency (IDF) measures the rarity of a word in the entire document set. The higher the TF-IDF value, the higher the importance of the word in the text and the more likely it is to be a keyword. For example, in multiple project information about electronic products, the word "smartphone" may appear frequently in one article (high TF value), but rarely in most other documents (high IDF value), so its TF-IDF value will be relatively high and it is very likely to be extracted as a keyword.

[0190] After that, use a rule-based method combined with the domain knowledge and business rules of the project information to formulate keyword extraction rules.

[0191] Then, perform part-of-speech tagging and screening. Perform part-of-speech tagging on the preprocessed text to identify words of different parts of speech such as nouns, verbs, and adjectives. Usually, nouns can better represent the core content of project information. Therefore, when extracting keywords, nouns can be preferentially selected. For example, for the sentence "This mobile phone has a high-definition camera function", nouns such as "mobile phone" and "camera function" can better reflect the key content of the project information and can be used as the key words to be extracted.

[0192] Keyword screening and sorting: After obtaining the candidate keywords according to the above process, screen the keywords according to the set threshold (such as the TF-IDF score threshold) to remove the words with lower scores;

[0193] At the same time, sort the screened keywords according to their importance scores, and select several keywords with higher rankings as the final project information keywords;

[0194] Perform project information duplicate checking and establish a project information database: Store the existing project information in the database, and each project information has a unique identifier;

[0195] The database structure should include the basic content of the project information, the extracted keywords, and other relevant attributes.

[0196] Establish an index for the keywords extracted from each project information in the database;

[0197] For quick searching and matching. An inverted index structure can be used, with the keywords as index items, and each keyword corresponds to a list of project information containing that keyword.

[0198] For the project information to be checked for duplicates, extract keywords according to the set keyword extraction rules, and match the keywords of the project information to be checked for duplicates with the keywords of the existing project information in the database;

[0199] Multiple matching strategies can be adopted, such as exact matching, fuzzy matching, etc. Exact matching means finding the project information in the database that is exactly the same as the keywords of the project information to be checked for duplicates; fuzzy matching allows a certain degree of difference. For example, by calculating the similarity between keywords (such as edit distance, cosine similarity, etc.) to determine whether there is similar project information.

[0200] Judge whether there is duplicate project information according to the matching results. If project information with highly matching keywords (exceeding the set similarity threshold) is found in the database for the project information to be checked for duplicates, it is considered a duplicate; otherwise, it is considered that there is no duplicate for this project information;

[0201] After all keyword matching is completed, generate duplicate checking information.

[0202] During the duplicate checking process in Step 4, when the imported project information is a bidding document, the specific duplicate checking process is as follows:

[0203] Import the bidding document into the preset standard form library for content reading:

[0204] Then, store the read content into the standard library line by line, paragraph by paragraph, or corresponding to specific fields according to the data entry specifications of the preset standard library. At the same time, record the basic information of the file, including the file name, upload time, and file source;

[0205] After that, perform a search for the compliant part. According to the preset standard rules, the standard rules exist in the metadata area of the standard library in the form of text descriptions, templates, and keyword sets;

[0206] For example, it is stipulated that certain specific chapters must contain certain key terms, specific format requirements (such as the row and column specifications of tables), etc.

[0207] After that, through text matching and pattern recognition technologies, screen out the paragraphs, clauses, and data items that conform to the standard rules in the content of the imported bidding document, mark them as compliant, and generate a search report to record the specific locations and content summaries that conform;

[0208] Then, extract the non-compliant part. Compare the content that has been retrieved as compliant, and use the difference comparison algorithm to find the parts that are different from it in the bidding document. The different parts include text expressions, numerical ranges, and format layouts;

[0209] Extract the different parts, that is, the difference parts, separately and organize them into an independent data set;

[0210] Attach the associated index of the original document to facilitate backtracking to find its source for subsequent analysis.

[0211] Input the extracted difference parts into a pre-trained deep learning model,

[0212] The architecture of the deep learning model is constructed based on convolutional neural networks (CNNs), recurrent neural networks (RNNs), and their variants, and is trained based on a large number of historical bidding documents and manually annotated standard data.

[0213] The deep learning model analyzes the difference parts from the aspects of semantic understanding, logical structure, and industry practices to judge whether they meet the requirements of industry general specifications, and outputs the analysis results. The analysis results include compliant and non-compliant;

[0214] For the difference parts with an analysis result of non-compliant, the system automatically extracts the difference features;

[0215] Such as incorrect term usage, missing required information, and illegal numerical settings, etc.

[0216] Record the differential features in a structured form and store them in the violation feature library of the preset standard library;

[0217] At the same time, establish associations with the corresponding bidding documents and the differential parts to facilitate subsequent statistical analysis and rule optimization.

[0218] When there is content that cannot be judged by the deep learning model, conduct manual judgment on this part of the content. When the manual judgment is non-compliant, extract the features of this part again and import them into the violation feature library of the preset standard library.

[0219] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered by the scope of the claims and the description of the present invention.

Claims

1. A project information analysis and duplication checking method, characterized by: The following steps are involved: Step 1: Import project information, which includes bidding documents and other project documents. Select different processing methods according to the type of project information to analyze the project information and convert it into a standard form. Step 2: After conversion to the standard form, the form is checked. If the form passes the check, semantic recognition is performed to obtain the analysis results of the filled content. If the form fails the check, a prompt message is generated to prompt the original project information to be checked. Step 3: When there is no abnormality in the analysis results, the project information analysis is carried out to obtain the project analysis results; Step 4: Analyze the project information, extract keywords, and check for duplicate project information after the extraction is completed; Step 5: After the project information is checked for duplicate content, the generated duplicate content information is sent to the preset receiving terminal.

2. The project information analysis and duplication checking method according to claim 1, characterized in that: The specific process in step one is as follows: Other types of project documents include paper handwritten project documents, paper printed project documents and electronic document version project documents; When the type of other project files is paper handwritten project files: First, the optical character recognition method based on deep learning is used to perform preliminary recognition to convert the handwritten text into preliminary recognition results; When the initial recognition results show suspected errors or unclear characters, the optical character recognition method based on semantic understanding is used for secondary recognition verification. The system automatically captures the text content before and after the character, combines the industry knowledge graph and semantic analysis algorithm to make corrections, and marks the corrected content. After obtaining the secondary recognition results, if there are no suspected errors or unclear characters, the next step is directly carried out. Finally, the manual verification module is used for manual inspection. If the manual verification passes, the final recognition result is exported; When the type of other project files is paper print project files: After scanning paper documents and converting them into digital images, a hierarchical optical character recognition operation is performed; The first layer uses a general optical character recognition engine to extract text content and identify the text body and basic format information; The second layer uses the intelligent optical character recognition module to recognize different elements such as titles, paragraphs, tables and charts using differentiated recognition strategies. Finally, a two-way format check is conducted to check whether the layout of each element in the digital image complies with the conventional format specifications of the industry from the perspective of visual presentation. If not, it will be rejected. From the perspective of text content logic, check whether the information between different elements is consistent, and finally export the recognition results that pass the two-way format check; When the type of other project files is electronic document version project files, directly convert them into standard formats; When the project information is a bidding document, the bidding document includes paper bidding documents and electronic bidding documents. For paper bidding documents, they are processed in the same way as paper-printed project documents, and for electronic bidding documents, they are directly converted into a standard format.

3. The project information analysis and duplication checking method according to claim 2 is characterized by: When the project information is other project files, the types of other project files include paper handwritten project files and paper printed project files. Before the paper handwritten project files are recognized, flatness detection is required. After the flatness detection passes, the next step of recognition is allowed.

4. The project information analysis and duplication checking method according to claim 3 is characterized by: The specific process of flatness detection is as follows: First, image acquisition is performed, that is, using an image acquisition device, image information of the paper handwritten project file is acquired at 45 degrees downward from the left, 45 degrees downward from the right, 45 degrees downward from the top, and 45 degrees downward from the bottom, to obtain first image information, second image information, third image information, and fourth image information; Then, extract the left edge center point T1, the right edge center point T2, the top edge center point T3 and the bottom edge center point T4 from the first image information; Then, the plane where the paper handwritten project file is placed is used as the reference plane, and the distance q1 between T1 and the reference plane, the distance m1 between T2 and the reference plane, the distance f1 between T3 and the reference plane, and the distance e1 between T4 and the reference plane are measured; Combine q1, m1, f1 and e1 to obtain the first evaluation parameter A1 (q1, m1, f1, e1); The second image information, the third image information and the fourth image information are processed in the same manner as the first image information to obtain the second evaluation parameter A2 (q2, m2, f2, e2), the third evaluation parameter A3 (q3, m3, f3, e3) and the fourth evaluation parameter A4 (q4, m4, f4, e4); Then, the difference Aa1 between the first evaluation parameter and the second evaluation parameter, and the difference Aa2 between the third evaluation parameter and the fourth evaluation parameter are calculated; Aa3 is a difference between the first evaluation parameter and the third evaluation parameter, Aa4 is a difference between the first evaluation parameter and the fourth evaluation parameter, Aa5 is a difference between the second evaluation parameter and the fourth evaluation parameter, Aa6 is a difference between the second evaluation parameter and the third evaluation parameter; When at least four of Aa1, Aa2, Aa3, Aa4, Aa5 and Aa6 exceed the preset range, it means that the flatness test fails.

5. The project information analysis and duplication checking method according to claim 4 is characterized by: The specific process of manual inspection through the manual proofreading module is as follows: When there is a second verification result, at least two people need to confirm the suspected wrong or unclear characters at least twice. If the two confirmation results are the same, the export is allowed; When the two confirmation results are different, re-calibration is performed.

6. The project information analysis and duplication checking method according to claim 1, characterized in that: The specific process of filling in the inspection is as follows: First, perform text area recognition, collect the filled-in image information, locate the position of the filled-in box in the image, and obtain its upper left corner coordinates (x1, y1) and lower right corner coordinates (x2, y2); Then, identify the range of the text in the fill-in box, and also obtain the coordinates of the upper left corner (x3, y3) and the lower right corner (x4, y4) of the text area; Then calculate the center coordinates of the text area, the horizontal coordinate x 中 =(x3+x4) / 2, ordinate y 中 =(y3+y4) / 2; Then calculate the center coordinates of the fill-in box, x 框中 =(x1+x2) / 2, ordinate y 框中 =(y1+y2) / 2; Calculate the horizontal coordinate deviation value Δx=|x 中 -x 框中 | and the vertical coordinate deviation value Δy=|y 中 -y 框中 |; When both the horizontal axis deviation value Δx and the vertical axis deviation value Δy are smaller than the set threshold, the text position is considered to be centered, indicating that the filling check has passed.

7. The project information analysis and duplication checking method according to claim 1, characterized in that: The project analysis results include projects with abnormalities and projects without abnormalities; The process of obtaining project analysis results is as follows: Extract the on-time completion rate of key nodes, average daily construction volume and the coordination of the progress of each subsystem from the project information; Process the on-time completion rate of key nodes, average daily construction volume and the coordination of the progress of each subsystem to obtain the project progress indicators; Then extract the proportion of batches that fail material quality inspection, rework rate and average failure interval of intelligent system from the project information; Process the proportion of unqualified batches in material quality inspection, rework rate and average failure interval of intelligent system to obtain quality control indicators; Then extract the cost-benefit ratio, budget change frequency and magnitude, and cost deviation rate from the project information; Process the cost-benefit ratio, budget change frequency and range, and cost deviation rate to obtain cost budget indicators; The project progress indicators, quality control indicators and cost budget indicators are processed to obtain comprehensive evaluation indicators. When the comprehensive evaluation indicator is less than the preset value, it means that the project is abnormal, otherwise it means that the project is normal.

8. The project information analysis and duplication checking method according to claim 7, characterized in that: The process of obtaining the project progress index is as follows: pre-establish a project progress score numerical mapping set of the on-time completion rate of key nodes, the average daily construction volume and the coordination of the progress of each subsystem, obtain the specific scores of the on-time completion rate of key nodes, the average daily construction volume and the coordination of the progress of each subsystem from the project progress score numerical mapping set, assign different weights to the on-time completion rate of key nodes, the average daily construction volume and the coordination of the progress of each subsystem, and then calculate the sum of the weighted scores to obtain the project progress index; The process of obtaining quality control indicators is as follows: A quality control score numerical mapping set of the proportion of batches that failed material quality inspection, the rework rate, and the average failure interval of the intelligent system is established in advance. After obtaining the specific scores of the proportion of batches that failed material quality inspection, the rework rate, and the average failure interval of the intelligent system from the control score numerical mapping set, different weights are assigned to the proportion of batches that failed building material quality inspection, the rework rate, and the average failure interval of the intelligent system, and then the sum of the weighted scores is calculated to obtain the quality control index; The process of obtaining cost budget indicators is as follows: A cost budget score numerical mapping set of cost-benefit ratio, budget change frequency and range, and cost deviation rate is established in advance. After obtaining the specific scores of cost-benefit ratio, budget change frequency and range, and cost deviation rate from the cost budget score numerical mapping set, different weights are assigned to the cost-benefit ratio, budget change frequency and range, and cost deviation rate, and then the sum of the weighted scores is calculated to obtain the cost budget indicator; The process of obtaining comprehensive evaluation indicators is as follows: Mark the project progress indicator as G1, the quality control indicator as G2, and the cost budget indicator as G3; Assign weight W1 to G1, weight W2 to G2, weight W3 to G3, W1+W2+W3=1, W3>W2=W1; Through the formula G1*W1+G2*W2+G3*W3=Gg, the comprehensive evaluation index Gg is obtained.

9. The project information analysis and duplication checking method according to claim 1, characterized in that: The checking process in step 4 is as follows: First, perform keyword extraction: clean the project information text, remove the noise data, and obtain the pre-processed thickness text; Then the importance score of each word in the project information text is calculated using the statistical TF-IDF method; Then, a rule-based approach is used to combine the domain knowledge and business rules to which the project information belongs to develop keyword extraction rules; Then perform part-of-speech tagging and screening, tag the preprocessed text with parts of speech, identify words with different parts of speech, and obtain candidate keywords; Keyword screening and sorting: After obtaining candidate keywords according to the above process, the candidate keywords are screened according to the set threshold, and words with scores less than the preset value are removed to obtain the screened keywords; At the same time, the filtered keywords are sorted according to the importance scores of the keywords, and several keywords with high rankings are selected as the final project information keywords; Check for duplicate project information and establish a project information database: store existing project information in the database, and each piece of project information has a unique identifier; Create an index for the keywords extracted from each project information in the database; For the project information to be checked for duplicates, keywords are extracted according to the set keyword extraction rules, and the keywords of the project information to be checked for duplicates are matched with the keywords of the project information already in the database; Determine whether there is duplicate project information based on the matching results. When project information with a similarity with the keywords of the project information to be checked for duplicates greater than a preset value is found in the database, it is considered that there is a duplicate; otherwise, it is considered that the project information is not duplicated.

10. The project information analysis and duplication checking method according to claim 1, characterized in that: During the duplication check in step 4, when the imported project information is a bidding document, the specific duplication check process is as follows: Import the bidding documents into the preset standard format library to read the content: Then, the read content is stored in the standard library line by line, paragraph by paragraph or according to specific fields according to the data entry specifications of the preset standard library, and the basic information of the file is recorded, including the file name, upload time and file source; After that, we will search for the parts that meet the standards. Based on the pre-set standard rules, we will use text matching and pattern recognition technology to select the paragraphs, clauses and data items that meet the standard rules from the imported bidding documents, mark them as meeting the standards, and generate a search report to record the specific locations and content summaries of the compliance. Then extract the parts that do not meet the standards, compare them with the retrieved content that meets the standards, and use the difference comparison algorithm to find out the parts that are different from the bidding documents; Extract the different parts, i.e. the difference parts, separately and organize them into independent data sets; The extracted differences are input into a pre-trained deep learning model. The deep learning model analyzes the differences from the perspective of semantic understanding, logical structure, and industry practices to determine whether they meet industry standards and outputs analysis results, including compliance and non-compliance. For the difference parts that do not meet the specifications in the analysis results, the system automatically extracts the difference features; Will The difference features are recorded in a structured form and stored in the violation feature library of the preset standard library; When the deep learning model cannot determine the content, the content will be manually judged. If it is manually judged as not meeting the standards, the features of the content will be extracted and imported into the violation feature library of the preset standard library.

Images