Construction engineering project file information management and storage method and system

Abstract

The invention relates to the technical field of data management, in particular to a construction engineering project file information management and storage method and system. According to the method, in an initial project category of a rough classification result according to project names, the name weighting difference degree which can pay more attention to key suffix file main bodies is calculated by setting position weights. The time difference degree and the department source difference degree serve as batch clustering features, engineering project files are clustered in sequence, and a final refined classification result is obtained. According to the method, the semantic analysis of position sensing and the time decay model are fused, so that high-precision automatic clustering and storage of the engineering project files are realized, and the automation level and filing accuracy of construction engineering file management are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of data management technology, specifically to a method and system for managing and storing construction project document information. Background Technology

[0002] With the digital transformation of the construction industry, engineering projects generate a massive amount of unstructured electronic documents throughout their entire lifecycle, covering various types such as planning permits, design drawings, construction records, and acceptance reports. These documents come from diverse sources, involving multiple responsible parties including construction companies, design institutes, contractors, and supervisors, and their number dynamically increases with the progress of the project. Currently, the management of such project documents mainly relies on manual classification and archiving or traditional retrieval systems based on keyword matching.

[0003] In practical engineering file management, existing classification technologies face serious challenges. First, engineering file names exhibit high homogeneity and repetition. Different projects or different stages of the same project often have numerous similar or even identical filenames such as "Design Specifications" and "Construction Schedule_V1". Traditional clustering methods based on text similarity (such as TF-IDF or ordinary word vectors) are often overly sensitive to common words in filenames, neglecting key distinguishing information such as version and specific sub-items usually included at the end of the filename. This leads to information mixing within the same engineering category, making subsequent data storage and updates inconvenient. Summary of the Invention

[0004] To address the problem that existing technologies suffer from poor classification and storage of engineering project files based solely on simple semantic naming, the present invention aims to provide a method and system for managing and storing construction project file information. The specific technical solution adopted is as follows: This invention proposes a method for managing and storing construction project document information, the method comprising: Obtain the file name, file generation timestamp, and file source department information for each project file in the initial project category of the database; obtain the word vector set for each project file based on the file name; obtain the department semantic feature vector based on the file source department information; For any two project files, the time interval between the file generation timestamps is compared with a preset time decay constant to obtain the time difference between the two project files; the position weight is obtained based on the position of the word vector in the word vector set; the difference between word vectors at the same position in the word vector sets of the two project files is weighted using the position weight to obtain the name weighted difference; the difference between the semantic feature vectors of the departments is used as the difference in the source of the departments. The time difference, name-weighted difference, and department source difference are used to cluster the project files in the initial project category in turn to obtain and store the file categories in the initial project category.

[0005] Furthermore, the method for obtaining the word vector set includes: The Word2Vec word vector model is used to process the file names into word vectors, resulting in multiple word vectors. The word vectors are arranged in the order of their appearance in the file names to obtain an initial word vector set. The number of word vectors for all project files is counted to obtain the maximum number of word vectors. If the number of word vectors for a project file is less than the maximum number of word vectors, a zero vector is added to the beginning of the initial word vector set until the number of vectors in the initial word vector set equals the maximum number of word vectors, thus obtaining the word vector set.

[0006] Furthermore, the method for obtaining the departmental semantic feature vector includes: The source department information of the document is processed using one-hot encoding to obtain the semantic feature vector of the department.

[0007] Furthermore, the method for obtaining the time difference includes: The ratio of the time interval to the preset time decay constant is used as the time difference degree.

[0008] Furthermore, the method for obtaining the position weights includes: Each vector is assigned a sequence number starting from 1 according to the order in which the vectors appear in the word vector set, and the ratio of the value corresponding to the sequence number to the maximum number of word vectors is used as the position weight.

[0009] Furthermore, the method for obtaining the name-weighted difference degree includes: The Euclidean distance between word vectors at the same position is used as the initial word vector difference. The initial word vector difference is weighted and averaged using the position weights to obtain the name weighted difference degree.

[0010] Furthermore, the method for obtaining the departmental source difference includes: The Euclidean distance between the semantic feature vectors of the departments is used as the source difference of the departments.

[0011] Furthermore, the process of clustering project files in the initial project category based on the time difference, name-weighted difference, and department source difference includes: Clustering is performed based on time differences to obtain the first cluster for different batches. Within each first cluster, clustering is performed based on name-weighted differences to obtain the second cluster. Within each second cluster, clustering is performed based on departmental origin differences to obtain the third cluster.

[0012] Furthermore, the DBSCAN clustering algorithm was selected as the clustering method.

[0013] The present invention also proposes a construction project document information management and storage system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of the construction project document information management and storage method described above.

[0014] The present invention has the following beneficial effects: This invention addresses the fact that the initial project category is already a coarse classification based on the project name. However, this coarse classification cannot further categorize projects based on their specific file types, such as acceptance reports and bidding proposals. Therefore, this invention obtains positional weights based on the position of word vectors within the set and weights the differences in word vectors, effectively distinguishing common prefixes from key suffixes in filenames. Compared to traditional text clustering, this method assigns higher weights to words in filenames that represent specific project sub-items or versions, effectively solving the misclassification problem caused by similar common filenames between different projects and improving the accuracy of filename similarity calculation.

[0015] Furthermore, by comparing the time interval between file generation timestamps with a preset time decay constant to obtain the time difference, this invention fully utilizes the "clustering" characteristic of engineering project files in terms of time. This mechanism can tightly aggregate files generated at similar times. Finally, clustering is performed sequentially according to the time difference, name-weighted difference, and department origin difference to obtain the final refined classification results. This invention achieves high-precision automatic clustering and storage of engineering project files by integrating location-aware semantic analysis and a time decay model, significantly improving the automation level and archiving accuracy of construction engineering file management. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for managing and storing construction project document information, as provided in one embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a construction project document information management and storage method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] The following description, in conjunction with the accompanying drawings, details the specific scheme of a construction project document information management and storage method and system provided by the present invention.

[0021] Please see Figure 1 The diagram illustrates a flowchart of a construction project document information management and storage method according to an embodiment of the present invention, the method comprising: Step S1: Obtain the file name, file generation timestamp, and file source department information for each project file in the database; obtain the word vector set for each project file based on the file name; obtain the department semantic feature vector based on the file source department information.

[0022] It should be noted that the embodiments of the present invention rely on the process of classifying only based on file names in the background technology. That is, before the implementation of the embodiments of the present invention, there are multiple initial project categories in the database. These initial project categories are classified based on the file names of project files. In the embodiments of the present invention, the word vectors corresponding to the file names of project files can be obtained using the Word2Vec word vector model. Then, the word vectors are converted into sentence vectors of file names using the Doc2Vec sentence vector model. For all project files, multiple initial project categories can be obtained by using the DBSCAN clustering algorithm through the Euclidean distance between sentence vectors. The embodiments of the present invention provide a more detailed division for each initial project category, so only one initial project category is used as an example for illustration. The language processing models and clustering methods mentioned above are all technical means well known to those skilled in the art, and will not be elaborated here.

[0023] For project documents requiring further classification, this specifically includes various design documents related to a particular project, such as design documents from various engineering construction units related to architecture, design, and construction, including design schemes for different stages of the project, meeting documents, test reports, record documents, etc., and the document creation time is recorded. Before storage, the file name, file creation timestamp, and document source department of each project document can be obtained. The resulting project documents to be stored may belong to the same project but not to the same stage; therefore, the purpose of this embodiment is to distinguish the stage and type of each project document within the initial project category and store them accordingly.

[0024] Therefore, this embodiment of the invention further obtains a set of word vectors for each engineering project file based on the file name; and obtains a semantic feature vector of the department based on the department information of the file source. Both the file name and the department information of the file source can be considered as objects of natural language processing. For example, an engineering project file might be named "Planning Report on Bridge Engineering," with the department information of the source department being the Housing and Construction Department, and the file generation timestamp accurate to the day.

[0025] Preferably, in this embodiment of the invention, the method for obtaining the word vector set includes: The Word2Vec word vector model is used to process the file name into word vectors, resulting in multiple word vectors. The word vectors are then arranged in the order of their appearance in the file name to obtain an initial set of word vectors.

[0026] Because subsequent steps require comparing word vectors at the same location in different project files, and the number of word vectors generated by different project files may differ, this embodiment of the invention considers that the targeted project files all have similar common prefixes, and the more important information lies in the key suffix names, such as "Planning Report on Project XX" and "Planning Blueprint for Project XX," where the planning report and planning blueprint are the key information components. Therefore, this embodiment of the invention counts the number of word vectors in all project files to obtain the maximum number of word vectors; if the number of word vectors in a project file is less than the maximum number of word vectors, then zero vectors are added to the beginning of the initial word vector set until the number of vectors in the initial word vector set equals the maximum number of word vectors, thus obtaining the word vector set. That is, by adding zero vectors to the beginning, the integrity of the later word vectors is ensured, facilitating subsequent comparative analysis.

[0027] Preferably, in this embodiment of the invention, the source department information of the document is processed using one-hot encoding to obtain the semantic feature vector of the department.

[0028] It should be noted that the natural language models and other semantic processing methods used above are all technical means well known to those skilled in the art, and will not be elaborated or limited here.

[0029] Step S2: For any two project files, compare the time interval between the file generation timestamps with the preset time decay constant to obtain the time difference between the two project files; obtain the position weight based on the position of the word vector in the word vector set; use the position weight to weight the difference of word vectors at the same position between the word vector sets of the two project files to obtain the name weighted difference; use the difference between the semantic feature vectors of the departments as the difference of the department source.

[0030] For a construction phase of an engineering project, the continuity of the project is relatively good, and the time intervals between the generated engineering documents are relatively close. Therefore, this embodiment of the invention first analyzes the time intervals between engineering project documents. Considering that for a fast-paced emergency repair project, a 3-day interval may indicate different phases, while for a large-scale infrastructure project spanning several years, a 30-day interval may still belong to the same batch, this embodiment of the invention not only quantifies the time difference based on the time interval, but also compares the time interval with a preset time decay constant to obtain the time difference. That is, the purpose of the preset time decay constant is to quantify the adaptability to the engineering project type corresponding to the current initial engineering category, converting the final time difference into a dimensionless difference.

[0031] Preferably, in this embodiment of the invention, the ratio of the time interval to the preset time decay constant is used as the time difference. In this embodiment of the invention, the preset time decay constant can be set to 7 days. If the time interval is less than the time decay constant, it indicates that the two files are strongly correlated in time, and the corresponding time difference is small; conversely, if the time decay is much greater than the time decay constant, it indicates that the two files may be historical files, not from the same batch, and the corresponding time difference is large.

[0032] This invention aims to achieve a detailed classification of documents within an initial project category based on batch, subject type, and department. The subject type information primarily includes suffixes in the document names, such as ****design drawings, ****meeting minutes, etc. Therefore, this invention further considers the position of word vectors within the word vector set to obtain positional weights. That is, the later the word vector appears, the greater its positional weight, and the more significant the difference reflected by the word vector at that position. Therefore, the positional weights are used to weight the differences in word vectors at the same position between the word vector sets of two project documents, obtaining a name-weighted difference degree. Finally, the difference between departmental semantic feature vectors is used as the departmental source difference degree, completing the analysis of various attribute information of the project documents.

[0033] Preferably, in this embodiment of the invention, since the number of word vectors in different project files has been standardized, the method for obtaining position weights includes: Each vector in the word vector set is assigned a sequence number starting from 1, and the ratio of the sequence number to the maximum number of word vectors is used as the position weight. Therefore, the method for obtaining the name-weighted difference includes: The Euclidean distance between word vectors at the same position is used as the initial word vector difference. The initial word vector difference is weighted and averaged using the position weights to obtain the name weighted difference degree.

[0034] It should be noted that the method for obtaining the Euclidean distance between vectors is a well-known technique in the art and will not be elaborated here.

[0035] Similarly, in this embodiment of the invention, the method for obtaining the departmental source difference includes: The Euclidean distance between the semantic feature vectors of the departments is used as the source difference of the departments.

[0036] Step S3: Cluster the project files in the initial project category according to the time difference, name weighted difference, and department source difference, respectively, to obtain and store the file categories in the initial project category.

[0037] Combining all the classification features calculated in the above steps, the project files in the initial project category can be clustered sequentially to obtain and store each file category in the initial project category. That is, the final result obtained in this embodiment of the invention is: in the disorganized initial project categories, multiple batch categories are divided according to time batches. In each batch, a project module is created for each project, and all project-related design drawings, project documents, meeting minutes, and other documents are stored in the corresponding project module folder for each project. Furthermore, departments belonging to the same category are further subdivided within these modules to complete a detailed classification, which is then finally stored in the database. This embodiment of the invention, based on the construction rules of building engineering and closely aligned with the construction process, organizes and stores the engineering data generated during construction, facilitating viewing and updating. This is beneficial for accelerating the implementation, follow-up, and management of special projects, ensuring project quality and safety, and making engineering projects more efficient and faster. It can effectively organize, store, and manage building engineering information, improve work efficiency, reduce risks, and ensure project quality and safety. In this field, automatic conflict detection and manual review functions can also be set simultaneously to optimize the iterative feedback mechanism of building document management. This process involves classifying and organizing construction project information documents, storing the classified documents in the relevant project modules of construction project information management, and placing them in the relevant database of construction project document management for easy access. The specific technical methods are well known to those skilled in the art and will not be elaborated here.

[0038] Preferably, in this embodiment of the invention, clustering is performed based on time difference to obtain first clusters of different batches. Within each first cluster, clustering is performed based on name-weighted difference to obtain second clusters. Within each second cluster, clustering is performed based on departmental origin difference to obtain third clusters. The clustering method used is the DBSCAN clustering algorithm. In each classification process, clustering is achieved by using the corresponding difference as the clustering distance. The specific clustering algorithm is a well-known technique to those skilled in the art and will not be elaborated upon here.

[0039] It should be noted that in other implementations of the present invention, the order of sequential clustering can also be changed to form a category topology structure suitable for its own needs, which will not be elaborated in detail.

[0040] In summary, this invention, in the initial project categories derived from coarse classification based on project name, calculates a weighted difference in name that better reflects the key file extensions by setting positional weights. The resulting time difference and departmental origin difference are used as features for batch clustering, and project files are clustered sequentially to obtain the final refined classification results. This invention, by integrating location-aware semantic analysis and a time decay model, achieves high-precision automatic clustering and storage of project files, significantly improving the automation level and archiving accuracy of construction project file management.

[0041] Based on the same inventive concept, the present invention also proposes a construction project document information management and storage system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the steps of the construction project document information management and storage method described above.

[0042] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0043] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for managing and storing construction project document information, characterized in that, The method includes: Obtain the file name, file generation timestamp, and file source department information for each project file in the initial project category of the database; obtain the word vector set for each project file based on the file name; obtain the department semantic feature vector based on the file source department information; For any two project files, the time interval between the file generation timestamps is compared with a preset time decay constant to obtain the time difference between the two project files; the position weight is obtained based on the position of the word vector in the word vector set; the difference between word vectors at the same position in the word vector sets of the two project files is weighted using the position weight to obtain the name weighted difference; the difference between the semantic feature vectors of the departments is used as the difference in the source of the departments. The time difference, name-weighted difference, and department source difference are used to cluster the project files in the initial project category in turn to obtain and store the file categories in the initial project category.

2. The method for managing and storing construction project document information according to claim 1, characterized in that, The methods for obtaining the word vector set include: The Word2Vec word vector model is used to process the file names into word vectors, resulting in multiple word vectors. The word vectors are arranged in the order of their appearance in the file names to obtain an initial word vector set. The number of word vectors for all project files is counted to obtain the maximum number of word vectors. If the number of word vectors for a project file is less than the maximum number of word vectors, a zero vector is added to the beginning of the initial word vector set until the number of vectors in the initial word vector set equals the maximum number of word vectors, thus obtaining the word vector set.

3. The method for managing and storing construction project document information according to claim 1, characterized in that, The method for obtaining the departmental semantic feature vector includes: The source department information of the document is processed using one-hot encoding to obtain the semantic feature vector of the department.

4. The method for managing and storing construction project document information according to claim 1, characterized in that, The method for obtaining the time difference includes: The ratio of the time interval to the preset time decay constant is used as the time difference degree.

5. The method for managing and storing construction project document information according to claim 2, characterized in that, The method for obtaining the position weights includes: Each vector is assigned a sequence number starting from 1 according to the order in which the vectors appear in the word vector set, and the ratio of the value corresponding to the sequence number to the maximum number of word vectors is used as the position weight.

6. The method for managing and storing construction project document information according to claim 5, characterized in that, The method for obtaining the name-weighted difference includes: The Euclidean distance between word vectors at the same position is used as the initial word vector difference. The initial word vector difference is weighted and averaged using the position weights to obtain the name weighted difference degree.

7. The method for managing and storing construction project document information according to claim 1, characterized in that, The methods for obtaining the departmental source difference include: The Euclidean distance between the semantic feature vectors of the departments is used as the source difference of the departments.

8. The method for managing and storing construction project document information according to claim 1, characterized in that, The process of clustering project files in the initial project category based on the time difference, name-weighted difference, and department source difference includes: Clustering is performed based on time differences to obtain the first cluster for different batches. Within each first cluster, clustering is performed based on name-weighted differences to obtain the second cluster. Within each second cluster, clustering is performed based on departmental origin differences to obtain the third cluster.

9. A method for managing and storing construction project document information according to claim 1, characterized in that, The DBSCAN clustering algorithm was selected as the clustering method.

10. A construction project document information management and storage system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the construction project document information management and storage method as described in any one of claims 1 to 9.

Images