Intelligent classification and grading method for building engineering data assets

By building a hierarchical classification directory and using large language models and computer vision technology to automatically classify and mark, combining the correlation score formula and asset score calculation, the problem of data classification chaos in construction engineering data asset management is solved, and the management of data assets is improved Efficiency and use value.

CN119088976AActive Publication Date: 2024-12-06SHANGHAI CONSTRUCTION FOURTH CONSTRUCTION GROUP CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411586309.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-08
Publication Date
2024-12-06
Estimated Expiration
2044-11-08

AI Technical Summary

Technical Problem

The lack of unified standards and specifications in the management of construction engineering data assets has led to confusion in data classification, and it takes a lot of time and manpower to find specific information, and it is impossible to make full use of the data value.

Method used

Using intelligent classification and grading methods, by building a hierarchical classification directory, using large language models and computer vision technology for document analysis and picture recognition, automatic document classification and marking, and automatically divide data assets through correlation scoring formulas and asset score calculations. Important level.

Benefits of technology

It realizes the precise classification and reasonable classification of construction engineering data assets, and improves the management efficiency and use value of data assets.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119088976A_ABST
    Figure CN119088976A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent classification and grading method for constructional engineering data assets, and belongs to the field of constructional engineering data asset management, and the method comprises the steps: collecting related data assets of constructional engineering, sorting the data assets, and constructing a hierarchical classification directory of the constructional engineering data assets; the content of the data assets is analyzed, and document classification marks are made; the data assets are divided into core assets, important assets and general assets according to the importance levels, the importance levels of a plurality of data assets are marked, and the importance levels of other data assets are automatically calculated. According to the method, the relevance formula and the asset scoring formula between the data assets of the components are utilized, the importance levels of other data assets can be automatically calculated, accurate classification and reasonable grading of the constructional engineering data assets are achieved, and the management efficiency and the use value of the data assets are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to an intelligent classification and grading method for construction engineering data assets, and belongs to the field of construction engineering data asset management. Background Art

[0002] With the rapid development of my country's construction industry, the number of construction projects has increased and the scale of projects has continued to expand, generating a large amount of construction project data assets. These data assets include design drawings, construction plans, bills of quantities, site photos, acceptance reports, etc., covering all stages of the project, including the early stage, construction process, and final settlement. These data assets in the construction industry are difficult to be efficiently managed and utilized.

[0003] The premise for the management and utilization of construction engineering data assets is to classify and grade these data assets, which are currently mainly processed manually. Due to the lack of unified standards and specifications, data asset classification is chaotic, and it is inconvenient to find and use; it takes a lot of time and manpower to find specific information in huge data assets; existing data asset management methods are difficult to discover differences in the importance of data and cannot fully utilize the value of data. Summary of the invention

[0004] In response to the problems existing in the existing construction project data asset management, the present invention provides an intelligent classification and grading method for construction project data assets, which can accurately classify and reasonably grade construction project data assets, thereby improving the management efficiency and use value of data assets.

[0005] In order to solve the above technical problems, the present invention includes the following technical solutions: An intelligent classification and grading method for construction engineering data assets, comprising: Step 1: Collect data assets related to construction projects, organize the data assets, and build a hierarchical classification catalog of construction project data assets; Step 2: Analyze the content of data assets and make document classification tags. For text data assets, use large language models to perform text analysis and automatically classify documents. For image data assets, use computer vision technology to identify and analyze engineering images, automatically classify engineering elements in the images, and perform document classification tags. Step 3: Classify data assets into core assets, important assets and general assets according to their importance level, mark the importance level of several data assets, and automatically calculate the importance level of other data assets, including: Step 3.1: Build any two data assets of the construction project P , Q Relevance score D The calculation formula is to determine P, Q Is it related? ; In the formula, L , k are the total number of levels and the k layer; I k is a 0-1 function, if P and Q Belong to the k Layer, then I k =1, otherwise I k =0; m It is extracted from step 2 P and Q The number of matches of document classification tags, s is the adjustment coefficient; Setting the Threshold D 0, if D > D 0, then determine that data assets < P , Q >Associated, otherwise, determine the data asset P , Q Not related; Step 3.2: Determine the relevance between data assets that are not labeled with importance and data assets that are labeled with importance; calculate asset scores S ,in, S =100 a +10 b + c , where a , b , c The number of unmarked importance assets associated with core assets, important assets, and general assets respectively; Step 3.3: Score by asset S Sort from large to small and mark the importance level of data assets according to set rules.

[0006] Furthermore, in step 3.3, the importance level of data assets is controlled in proportion and the asset scores are S After sorting from largest to smallest, C 1% of the scored data assets are marked as core assets. C 1% to C 2% of data assets are marked as important assets. C Data assets after 2% are marked as general assets. C 1. C 2. C3 is the default value, and 0< C 1< C 2<100.

[0007] further, C 1=5, C 2=20.

[0008] Furthermore, the hierarchical classification catalog of construction engineering data assets includes six layers, specifically: The first level: classification according to the data generation stage; The second level: classified by data source; The third layer: classification by data usage; The fourth layer: classification according to the nature of data; The fifth layer: classified by application scenarios; Layer 6: Classification according to business needs.

[0009] Furthermore, a hierarchical clustering algorithm is used to intelligently distribute the data assets collected in step one into a hierarchical classification directory of construction engineering data assets.

[0010] Furthermore, σ=2.

[0011] Further, in step 3.1, D 0 is obtained in the following way: the total number of data assets is N, all data assets are randomly paired, and the correlation scores are calculated, all correlation scores are sorted, and the 90% quantile is statistically taken as the threshold D 0.

[0012] Due to the adoption of the above technical scheme, the present invention has the following advantages and positive effects compared with the prior art: the intelligent classification and grading method for construction project data assets provided by the present invention first constructs a data asset hierarchical classification directory based on the collected construction project related data assets, allocates the data assets to the data asset hierarchical classification directory, and also analyzes the content of the data assets and makes document classification marks. After marking the importance levels of some data assets, the importance levels of other data assets can be automatically calculated by using the correlation formula and asset scoring formula between data assets, thereby realizing accurate classification and reasonable grading of construction project data assets and improving the management efficiency and use value of data assets. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flowchart of a method for intelligent classification and grading of construction project data assets in one embodiment of the present invention. DETAILED DESCRIPTION

[0014] The following is a further detailed description of the intelligent classification and grading method for construction engineering data assets provided by the present invention in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become clearer in conjunction with the following description. It should be noted that the accompanying drawings are all in a very simplified form and are not in precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0015] This embodiment provides a method for intelligent classification and grading of construction project data assets, which specifically includes the following steps: Step 1: Collect construction project-related data assets, organize the data assets, and build a hierarchical classification catalog of construction project data assets.

[0016] The collected construction project-related data assets include design drawings, construction plans, bills of quantities, on-site photos, etc., and the collected data assets are pre-processed, such as cleaning, deduplication and format unification.

[0017] The data asset classification directory can be pre-set to divide the data assets into L For example, the first layer includes 3 categories, and the second layer includes 4 categories. The 4 categories in the second layer are obtained by further dividing each of the 3 categories in the first layer into 4 categories.

[0018] The collected data assets need to be allocated into a hierarchical classification directory of data assets, and a hierarchical clustering algorithm can be used for intelligent allocation.

[0019] In a specific embodiment, the hierarchical classification directory of construction engineering data assets includes six layers, as follows: The first level: classified according to the stage of data generation, for example, divided into: planning stage, design stage, bidding stage, construction stage, and completion stage.

[0020] The second level: classified by data source. For example, it can be divided into: provided by Party A, provided by the design unit, provided by the construction unit, provided by the supervision unit, and provided by a third-party testing agency.

[0021] The third layer: Classification by data usage. For example, it can be divided into: design data, construction planning, cost control, quality and safety management, and project site management.

[0022] The fourth layer: Classification by data nature, for example, text, drawings, images, and databases.

[0023] The fifth layer: classified by application scenario. For example, it can be divided into: design scenario, construction scenario, supervision scenario, and construction scenario.

[0024] The sixth layer: classified according to business needs. For example, it can be divided into: engineering entity needs, safety needs, and environmental protection needs.

[0025] It should be noted that the specific number of layers of construction project data assets and the specific categories of each layer can be set as needed.

[0026] Step 2: Analyze the content of data assets and make document classification tags.

[0027] Document classification tagging is to tag a specific data asset based on its content, that is, to set a number of tags based on the identified content. These tags are equal in status and do not involve a hierarchical relationship. Their main function is to determine the number of tag matches between two data assets in step 3. m . For example, for text data assets, large language models can be used to perform text analysis on engineering materials and automatically classify and mark documents; for image data assets, computer vision technology can be used to identify and analyze engineering images and automatically classify engineering elements in the images. For example, documents can be marked as text, images, drawings, databases, etc. according to their content and format; text data can be identified and stage classifications can be marked according to the construction stages involved in the document content, such as planning stage, design stage, bidding stage, construction stage, completion stage, etc.; topic classifications can be marked according to the professional fields of the document content, such as structural engineering, electromechanical installation, decoration and renovation, gardening and landscape, municipal engineering, etc.; keywords in documents can be extracted through natural language processing technology, such as concrete, steel structure, energy-saving sub-projects, etc. Using computer vision technology, we can identify elements such as buildings, equipment, and materials in the image, and mark them as buildings, tower cranes, building materials (rebars), etc. according to the engineering elements; mark the construction status according to the recognition results and classify them into basic construction, main construction, and decorative construction; mark quality problems according to the recognition results and classify them into cracks, leakage, and deformation; and classify the scenes according to the recognition demerits into construction sites, material yards, and office areas. The marking type can be consistent with the name of the hierarchical classification, or have a corresponding relationship with the name of the hierarchical classification, so as to facilitate the allocation of data assets to the hierarchical classification table.

[0028] Step 3: Classify data assets into core assets, important assets and general assets according to their importance levels, mark the importance levels of several data assets, and automatically calculate the importance levels of other data assets.

[0029] All data assets with marked importance levels are marked as classified, and other data assets are marked as unclassified. Specifically: according to the characteristics of construction projects, rules for distinguishing important assets are formulated, such as key processes, important parts, high-value assets, etc.; several data assets are marked for each importance level.

[0030] Automatically calculate the importance level of other data assets, including: Step 3.1: Build two data assets for construction projects P , Q Relevance score D The calculation formula is to determine P , Q Is it related? ; In the formula, L , k are the total number of levels and the k layer; I k is a 0-1 function, if P and Q Belong to the k Layer, then I k =1, otherwise I k =0; m It is extracted from step 2 P and Q The number of matches of document classification tags, s is the adjustment coefficient, preferably σ=2.

[0031] As an example, P and Q The classification is the same in the first to fourth layers, but different in the fifth layer, and naturally different in the sixth layer, and P and Q There are 11 identical tags, which are matches. So: D =1×2 1 +1×2 2 +1×2 3 +1×2 4 +0×2 5 +0×2 6 +2×11=52.

[0032] Setting the Threshold D 0, if D > D 0, then the two assets are considered to be related. D 0 can be preset based on experience. In a specific embodiment, D 0 is obtained in the following way: the total number of data assets is N, all data assets are randomly paired, and the correlation scores are calculated, all correlation scores are sorted, and the 90% quantile is statistically taken as the threshold D 0.

[0033] Step 3.2: Determine the relevance between data assets that are not labeled with importance and data assets that are labeled with importance; calculate asset scores S,in, S =100 a +10 b + c , where a , b , c It is the number of unmarked importance assets associated with core assets, important assets, and general assets respectively.

[0034] Step 3.3: Score by asset S Sort from large to small, and set the top 5% of the scores as core assets, 5% to 20% as important assets, and above 20% as general assets.

[0035] If a new data asset enters the classification system, set the asset as "unclassified" and determine the importance level of the data asset according to steps 3.2 and 3.3.

[0036] This embodiment divides the importance level of data assets by controlling the percentage of quantity. Through steps 3.2 and 3.3, the unclassified data assets are entered into the classified ones. At this time, the total number of the classified data assets changes. The asset scores of all data assets can be calculated again and re-sorted to obtain the importance level of each data asset again.

[0037] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described 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.

[0038] The above-mentioned embodiments only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A method for intelligent classification and grading of construction engineering data assets, characterized in that: include: Step 1: Collect data assets related to construction projects, organize the data assets, and build a hierarchical classification catalog of construction project data assets; Step 2: Analyze the content of data assets and make document classification tags. For text data assets, use large language models to perform text analysis and automatically classify documents. For image data assets, use computer vision technology to identify and analyze engineering images, automatically classify engineering elements in the images, and perform document classification tags. Step 3: Classify data assets into core assets, important assets and general assets according to their importance level, mark the importance level of several data assets, and automatically calculate the importance level of other data assets, including: Step 3.1: Build any two data assets of the construction project P , Q Relevance score D The calculation formula is to determine P , Q Is it related? ; In the formula, L , k are the total number of layers and the k layer; I k is a 0-1 function, if P and Q Belong to the k Layer, then I k =1, otherwise I k =0; m It is extracted from step 2 P and Q The number of matches of document classification tags, σ is the adjustment coefficient; Setting the Threshold D 0, if D > D 0, then determine that data assets < P , Q >Associated, otherwise, determine the data asset P , Q Not related; Step 3.2: Determine the relevance between data assets that are not labeled with importance and data assets that are labeled with importance; calculate asset scores S ,in, S =100 a +10 b + c , where a , b , c The number of unmarked importance assets associated with core assets, important assets, and general assets respectively; Step 3.3: Score by asset S Sort from large to small and mark the importance level of data assets according to set rules.

2. The intelligent classification and grading method for construction engineering data assets according to claim 1, characterized in that: In step 3.3, the importance level of data assets is controlled in proportion and the asset scores are S After sorting from largest to smallest, C 1% of the scored data assets are marked as core assets. C 1% to C 2% of data assets are marked as important assets. C Data assets after 2% are marked as general assets. C 1. C 2. C 3 is the default value, and 0< C 1< C 2<100.

3. The intelligent classification and grading method for construction engineering data assets according to claim 2 is characterized in that: C 1=5, C 2=20。 4. The intelligent classification and grading method for construction engineering data assets according to claim 1, characterized in that: The hierarchical classification catalog of construction engineering data assets includes six layers, specifically: The first level: classification according to the data generation stage; The second level: classified by data source; The third layer: classification by data usage; The fourth layer: classification according to the nature of data; The fifth layer: classified by application scenarios; Layer 6: Classification according to business needs.

5. The intelligent classification and grading method for construction engineering data assets according to claim 4 is characterized in that: A hierarchical clustering algorithm is used to intelligently allocate the data assets collected in step one to a hierarchical classification directory of construction engineering data assets.

6. The intelligent classification and grading method for construction engineering data assets according to claim 1, characterized in that: σ=2.

7. The intelligent classification and grading method for construction engineering data assets according to claim 1, characterized in that: In step 3.1, D 0 is obtained in the following way: the total number of data assets is N, all data assets are randomly paired, and the correlation scores are calculated, all correlation scores are sorted, and the 90% quantile is statistically taken as the threshold D 0.

Citation Information

Patent Citations

  • Network key asset identification method

    CN110460481A

  • Intelligent classification and grading system for information asset visual management method library

    CN114707599A

  • Data asset classification and grading method, system, equipment and medium

    CN116401321A

  • Data self-service asset directory construction method and device based on machine learning

    CN116932550A

  • Data asset management method and system

    CN118897908A