LU504881B1

LU504881B1 - Intelligent collection method and system for engineering archives based on enabling thinking

Info

Publication number: LU504881B1
Application number: LU504881A
Authority: LU
Inventors: Hui Chen
Original assignee: Univ Central China Normal
Priority date: 2023-08-08
Filing date: 2023-08-08
Publication date: 2024-02-08

Abstract

The invention discloses an intelligent collection method and system for engineering archives based on enabling thinking, including six steps of intelligent collection, intelligent classification, intelligent identification, intelligent generating test paper, intelligent cataloging, and intelligent visa. The step of intelligent collection includes automatic reception, intelligent audit, and automatic registration; the step of intelligent classification includes the classification of the category and the determination of the classification number; the step of intelligent identification includes storage period division and security level identification; the step of intelligent generating test paper include intelligent assistance of generating test paper, case paper sorting, and in-volume file sorting, the step of intelligent cataloging include automatically generating page numbers and file numbers; the step of intelligent visa steps include intelligent visa processing and intelligent signature identification. This method is mainly oriented to the intellectualization of engineering archives collection.

Description

DESCRIPTION 17006981

INTELLIGENT COLLECTION METHOD AND SYSTEM FOR ENGINEERING

ARCHIVES BASED ON ENABLING THINKING

TECHNICAL FIELD

[01] The invention relates to the technical field of archives management, in particular to an intelligent method and system for collecting engineering archives based on enabling thinking.

BACKGROUND ART

[02] As an important information resource supporting engineering construction, engineering archives are the key components in the whole project management process. Under the digital environment, the number of engineering archives has increased dramatically, and the types of engineering archives resources are more diverse. The manual archives collection method reveals problems such as cumbersome procedures, low efficiency, repetitive labor, and omissions.

[03] With the increasing prominence of these problems, it is more urgent to adopt intelligent means to break through the dilemma of traditional methods of archives consolidation and ensure the procedural, accurate, and efficient needs of archives consolidation. Therefore, various intelligent Management systems or platforms are increasingly and widely used as effective tools for archives consolidation.

[04] Due to the single function, the existing intelligent management system can neither excavate the integration effect between multiple functions nor realize the intelligentization of the whole process of archives collection.

[05] From the perspective of platform construction of the system, most of the existing implementation schemes only use traditional technical means and still rely on manual processing to a large extent.

The existing file management system or method mainly applies scanning 7504881 technology, sensing technology, encryption technology, and RFID technology, the level of automation and intelligence is relatively low.

[06] The existing intelligent scheme only constructs the scheme for a certain link of archives management, it does not realize the intelligent scheme design of the whole process of archives collection.

[07] From the perspective of the industry, the scheme of applying technical means to carry out archives management involves industrial and commercial digital archives, accounting electronic archives, urban construction archives, and other fields, But those fields can not be well adapted to the characteristics of large volume, many types and wide subjects of engineering project archives, and the actual needs of engineering project archives can not be satisfied. Therefore, there is a certain gap in the implementation scheme in the field of intelligent integration of engineering archives.

[08] As a complete process, engineering archives work data comes from different business systems, so it is easy to produce data heterogeneity problems in collaborative cooperation. And the application of artificial intelligence technology in the existing business systems is in the primary stage, and the level of intelligence still needs to be improved.

SUMMARY

[09] The purpose of this invention is to provide an intelligent collection method and system for engineering archives based on enabling thinking. This method is mainly oriented to the intelligentization of engineering archives collection. By mining practical problems and analyzing key requirements, it mainly integrates data, technology, and knowledge to ensure the standardization of engineering archives collection and effectively improve the collection efficiency of engineering archives.

[10] In order to achieve the above purpose, the invention provides an 7504881 intelligent collection method for engineering archives based on enabling thinking, this method includes the following steps:

[11] Step 1: intelligent collection, completing an automatic reception, audit, and registration of other business system interface archiving files intelligently through an application of artificial intelligence algorithms;

[12] Step 2: intelligent classification, processing, and analyzing text contents of engineering archives with the help of an artificial intelligence algorithm to divide attribution categories and determine classification numbers automatically;

[13] Step 3: intelligent identification, extracting intelligent classification results and forming a rule set with the help of an artificial intelligence algorithm, and then calling the results and the rule set through a rule engine to divide attribution categories and determine classification numbers automatically;

[14] Step 4: intelligent generating test paper, simulating the process of generating and arranging physical files in reality, completing intelligent assistance of generating test paper, file sorting, and in-volume file sorting;

[15] Step 5: intelligent cataloging, realizing two independent functions of page number automatic writing and file number automatic generation by integrating deep learning and natural language processing algorithms;

[16] Step 6: intelligent visa, introducing artificial intelligence algorithm to realize the intelligentization of the visa process of the completion file.

[17] Preferably, Step 1 also includes the following steps:

[18] Step 1.1: adding an automatic receiving program to perform preprocessing;

[19] Step 1.2: identifying PDF, ODF format text files, photo files, and other multimedia files by an algorithm, and then using an information extraction algorithm to analyze the morphology and syntax of the identified information;

[20] Step 1.3: entering an intelligent audit link, and eliminating files that do not meet the requirements;

[21] Step 1.4: starting a registration process, giving an electronic file a unique 7504881 identifier automatically, using a natural language processing algorithm to extract the form entry, and using a feature matching algorithm to match the file content and the information entry to realize the automatic filling of the information of a registration form.

[22] Preferably, Step 2 also includes the following steps:

[23] Step 2.1: applying a recognition algorithm to identify registered engineering project documents and multimedia electronic archives;

[24] Step 2.2: using a natural language processing algorithm to analyze a basic analysis, an attribute analysis, a semantic analysis, and a structural analysis of an identified engineering archives, finding a relationship model between a file text attribute and a file archiving category according to a pre-set archiving template;

[25] Step 2.3: sampling scanned parts in an existing engineering archives collection, cleaning the data according to input requirements of a learning algorithm, constructing a training set and a testing set, and achieving an expected learning effect by constantly improving the model;

[26] Step 2.4: inputting registered project files into a relational model, matching the files to a filing scope in a filing module, and further refining a filing category to realize the intelligent classification of a file ownership category in the whole file;

[27] Step 2.5: on the basis of accurate classification of file categories, matching a classification number set by an archiving template according to the second-level category name, and supporting an authorized user to set the classification code for the newly created files.

[28] Preferably, Step 3 also includes the following steps:

[29] Step 3.1: creating a rule engine and determine an interface to invoke a rule set;

[30] Step 3.2: compiling and loading external business rules into a rule set for the rule engine to call and execute;

[31] Step 3.3: extracting results of intelligent classification as a set of data objects processed by rule sets and adding the set to a working memory; 7504881

[32] Step 3.4: executing a rule matching command by an engine, matching an application object in the working memory with a condition part of the rule in a rule set container, returning a corresponding value of a classification number in an archive template and a storage period table, and deriving a preliminary judgment result.

[33] Preferably, Step 4 also includes the following steps:

[34] Step 4.1: inputting identified files into a machine learning model;

[35] Step 4.2: incorporating set rules of file arrangement and generating test paper in s volume, operation norms, and experience summary in the practice of arrangement and generating test paper into the rule base of deep learning to support the deep learning model;

[36] Step 4.3: training and modeling the system by using a file set of sorted and grouped papers to form a computer-readable sorting and a grouped paper rule;

[37] Step 4.4: identifying the nature of the file through a sorting law of the files in the volume, and clustering according to the same characteristics to assist generating test papers; extracting topic keywords automatically through a deep learning algorithm, and generating a file title and the key information of the file in combination with title generation rules provided;

[38] Step 4.5: sorting files and files in the volume according to file sorting rules, and generating the management of the files and files in the volume automatically.

[39] Preferably, Step 6 also includes the following steps:

[40] Step 6.1: uploading archives that need to be reviewed by personnel of all parties according to the requirements of visa processing;

[41] Step 6.2: receiving visa information by the system automatically, conducting intelligent audits of visa information, and uploading data through intelligent audit algorithms;

[42] Step 6.3: introducing the image recognition algorithm to extract the signatures involved in the file visa, and standardize the entry of signature data through data preprocessing; 7504881

[43] Step 6.4: extracting global features and local features of the signature data after preprocessing, and comparing them with the features of a registered signature and a verification signature, calculating the feature similarity, and outputting a signature identification result according to a set feature similarity preset value;

[44] Step 6.5: transferring the final results of an intelligent audit to an interface of a corresponding audit unit and department.

[45] Preferably, in step 1.3, the intelligent audit link also includes a determination of the filing scope, a quality audit, and a batch check.

[46] An intelligent collection system for engineering archives based on enabling thinking, this system includes an intelligent collection module, an intelligent classification module, an intelligent identification module, an intelligent generating test paper module, an intelligent cataloging module, and an intelligent visa module.

[47] A computer readable storage medium that stores computer programs, the method described in any of claims 1 to 7 is implemented when the computer program is executed by a processor.

[48] The advantages and positive effects of the intelligent method and system for collecting engineering archives based on enabling thinking described in the invention are as follows:

[49] The invention realizes the integration of the technical ability of artificial intelligence and the resource integration ability of digital platform, which is convenient for applying artificial intelligence technology to solve stylized and repetitive file collection and collation business, so as to realize the intelligent integration of engineering file collection and collation work and improve the efficiency and level of file collection and collation.

[50] The invention comprehensively considers the six steps in the whole 7504881 process of archives from collection to visa collection, and the new technology is applied to the overall business process of archives, it pays attention to the diversity and systematicness of functions, it is conducive to multi-functional interaction and exerts the positive synergistic effect of '1+1>2".

[51] The invention focuses on the in-depth application of various advanced intelligent technologies such as pattern recognition, natural language processing, and deep learning in the workflow of engineering archives consolidation, and combines the technical characteristics with the characteristics of engineering project archives, which can better improve the efficiency and level of intelligent consolidation of engineering archives.

[52] The following is a further detailed description of the technical solution of the invention through drawings and an embodiment.

BRIEF DESCRIPTION OF THE DRAWINGS

[53] Fig. 1 is a method flow chart of the invention;

[54] Fig. 2 is a collection flow chart of Step 1;

[55] Fig. 3 is a classification flow chart of Step 2;

[56] Fig. 4 is an identification flow chart of Step 3;

[57] Fig. 5 is a flow chart of Step 4;

[58] Fig. 6 is a visa flow chart of Step 6.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[59] The following is a further explanation of the technical solution of the invention through drawings and an embodiment.

[60] Embodiment

[61] As shown in Fig. 1, an intelligent collection method for engineering archives based on enabling thinking, this method includes the following steps:

[62] Step 1: intelligent collection, completing an automatic reception, audit, 7504881 and registration of other business system interface archiving files intelligently through an application of artificial intelligence algorithms, as shown in Fig. 2.

[63] Step 1.1: adding an automatic receiving program to realize batch intelligent automatic receiving of archive files of other business system interfaces, and performing preprocessing (processing missing values, inconsistent data formats, etc.); at this time, a view is used to control the user's authority, so that the department's part-time archivists can only see the archived files of the department; archivists can see the archived files of all departments, monitor the collation of archived files of all departments, and receive or call back the archived files collated by the department. The collection module is a public module provided to file managers and departmental part-time archivists for archiving documents.

[64] Step 1.2: identifying text files, photo files, and other multimedia files in

PDF and ODF formats through OCR recognition, speech recognition, image recognition, and other pattern algorithms, and then analyzing the morphology and syntax of the identified information with the help of information extraction algorithm, so the comprehensive extraction of text content, text structure, and metadata is realized, and the empowerment of intelligent algorithms and archival data resources is realized. Taking the text file as an example, the OCR recognition algorithm automatically identifies the scanned file and extracts the text content, structure, and metadata such as ‘title’, 'sub-title', 'subject word’, ‘keyword’, ‘responsible person’, 'date’, language’, 'main delivery", 'copy delivery’, ‘number of pieces’, and 'number of pages' for preliminary information extraction.

In the identification of scanned electronic documents, through its layout analysis function, it can not only realize the scanning and identification of documents quickly, but also extract the key information of the files automatically, and the information is filled in the directory information of the files after automatic identification, which reduces the entry workload greatly, and the possibility is provided for batch entry of document files.

[65] Step 1.3: entering the intelligent audit link, reviewing whether the 7504881 electronic file meets the filing scope and whether there is a quality problem, and eliminating files that do not meet the requirements; the intelligent audit process also includes the determination of the scope of filing, quality audit, and batch verification.

[66] A project file archiving template shown in Table 1, for the determination of the archiving scope, it is necessary to match the extracted key text information with the specified project file archiving template. The files with successful archiving scope matching are then based on the previously extracted information items. The similarity calculation is carried out, and the files with extremely high similarity are eliminated, and the files that meet the archiving scope are left.

[67] For quality audit, a rule engine is used to intelligently check the extracted text information with the specified engineering file acceptance inspection standards, audit points, and archiving templates, and quality inspection is performed on the standardization of file compilation, content accuracy, and quantity completeness.

[68] In the process of batch verification, the above standards and rules are disassembled into a fixed number of filtering rules by the computer system, which is made into a check template that can be recognized by the computer, and batch verification is realized by matching the template with the information.

[69] Step 1.4: starting a registration process, giving an electronic file a unique identifier automatically, using a natural language processing algorithm to extract the form entry, and using a feature matching algorithm to match the file content and the information entry to realize the automatic filling of the information of a registration form.

[70] Step 2: intelligent classification, processing, and analyzing text contents of engineering archives with the help of artificial intelligence algorithms such as natural language processing and deep learning to divide attribution categories and determine classification numbers automatically; as shown in Fig. 3.

[71] Step 2.1: applying OCR, image and speech recognition, and other pattern 7504881 recognition algorithms to identify the registered project documents and multimedia electronic files, such as construction photos and completion drawings.

[72] Step 2.2: using a natural language processing algorithm to analyze a basic analysis, an attribute analysis, a semantic analysis, and a structural analysis of an identified engineering archives, finding a relationship model between a file text attribute and a file archiving category according to a pre-set archiving template; For example, after a text file is identified, its continuous natural language text is cut into a lexical sequence with semantic rationality and integrity, and the words in its language are part-of-speech tagging, that is, attribute analysis, indicating verbs, nouns and adverbs, etc., and the identification of proper nouns, including names of people, institutions and places, in order to distinguish the theme. Dependency syntax analysis is performed on a large number of unstructured texts, from which information such as entities, concepts, and semantic relationships are extracted to construct domain knowledge. On the basis of the above content, the relationship analysis is used to carry out synonym association and semantic web association on the core concepts in the text files, and match the similar file types to realize the file classification.

[73] Step 2.3: sampling scanned parts in an existing engineering archives collection, cleaning the data according to input requirements of a learning algorithm, constructing a training set and a testing set, and achieving an expected learning effect by constantly improving the model;

[74] Step 2.4: inputting registered project files into a relational model, matching the files to a filing scope in a filing module, and further refining a filing category to realize the intelligent classification of a file ownership category in the whole file, including an intelligent division of the first-level category and the second-level category, that is, all files from a specific institution (including administrative agencies, companies, institutions, individuals, etc.).

[75] Step 2.5: on the basis of accurate classification of file categories, matching a classification number set by an archiving template according to the second-level category name, and supporting an authorized user to set the 7504881 classification code for the newly created files.

[76] Step 3: intelligent identification, extracting intelligent classification results and forming a rule set with the help of an artificial intelligence algorithm, and then calling the results and the rule set through a rule engine to divide attribution categories and determine classification numbers automatically, as shown in Fig. 4.

[77] Step 3.4: returning a corresponding value of a classification number in an archive template and a storage period table, and deriving a preliminary judgment result.

[78] Step 3.1: creating a rule engine and determine an interface to invoke a rule set; for example, the custody period of a document file is 25 years, and the document files to be archived include bidding announcements, bidding applications, etc.

[79] Step 3.2: compiling and loading external business rules such as the project file archiving template and the retention period table shown in Table 1 into a rule set for the rule engine to call and execute.

[80] Step 3.3: extracting results of intelligent classification as a set of data objects processed by rule sets and adding the set to a working memory; the results of the classification of the above files are put into the memory as the data object of the identification.

[81] Step 3.4: executing a rule matching command by an engine, matching an application object in the working memory with a condition part of the rule in a rule set container, and selecting a feature matching algorithm when matching, the first step is to find out the content with strong characteristics in the file and match it with the content in the rule set container by extracting its key content. In this process, different feature similarity calculation methods can be used according to different matching objects. Using a cosine distance calculation formula:

oo LU504881

[82] het i He

[83] For example, according to the rule set, the conformities of the two files of the source text and the text to be identified by the two machines are (X1, Y1)(Xz,

Y2), respectively, the X and Y corresponding vectors of the extracted object are calculated to compare the similarity. When the cosine value is smaller, the two elements are less relevant.

[84] The Euclidean distance is to extract the three-dimensional spatial features of an object and match its features to analyze the differences in individual numerical features.

[85] The Jaccard similarity is mostly used to determine whether the object has a certain feature element. By analyzing the matching degree of different sample objects on the elements, the similarity between the two is analyzed. Let sample

A and sample B be two n-dimensional vectors, and the values of all dimensions are 0 or 1. For example, A (0,1,1,0) and B (1,0,1,1). We regard the sample as a set, 1 means that the set contains the element, and 0 means that the set does not contain the element. Finally, after similarity calculation, the corresponding values of the classification number in the archive template and the storage period table are returned, and the preliminary judgment results are derived, such as the storage period of 10 years, 25 years, and permanent; the secret level is top secret, confidential, secret, public.

[86] For the archives with storage period and secret level in advance, the intelligent identification results are compared with the original records. If the comparison results are inconsistent, the original storage period and secret level identification results will be marked.

[87] Step 4: intelligent generating test paper, simulating the process of generating and arranging physical files in reality, the computer completes intelligent assistance of generating test paper, file sorting, and in-volume file sorting; as shown in Fig. 5, the three functions of the module are all based on deep learning algorithms and need to be implemented through models.

[88] Step 4.1: inputting identified files into a machine learning model; 7504881

[89] Step 4.2: incorporating set rules of file arrangement and generating test paper in s volume, operation norms, and experience summary in the practice of arrangement and generating test paper into the rule base of deep learning to support the deep learning model;

[90] Step 4.3: training and modeling the system by using a file set of sorted and grouped papers to form a computer-readable sorting and a grouped paper rule;

[91] Step 4.4: identifying the nature of the file through a sorting law of the files in the volume, such as the management of the text engineering algorithmic files, etc., and clustering according to the same characteristics to assist generating test papers, such as the same theme, the same storage period, etc.; extracting topic keywords automatically through deep learning of the machine, and generating a file title and the key information of the file in combination with title generation rules provided;

[92] Step 4.5: sorting files and files in the volume according to file sorting rules, and generating the management of the files and files in the volume automatically.

[93] Step 5: intelligent cataloging, realizing two independent functions of page number automatic writing and file number automatic generation by integrating deep learning and natural language processing algorithms; The specific process is similar to that of intelligent generating test paper, and the main difference is reflected in business rules.

[94] Step 6: intelligent visa, introducing artificial intelligence algorithm to realize the intelligentization of the visa process of the completion file; for visa intelligent processing and signature intelligent identification, the specific steps are shown in Fig. 6.

[95] Step 6.1: uploading archives that need to be reviewed by personnel of all parties according to the requirements of visa processing;

[96] Step 6.2: receiving visa information by the system automatically, and 7504881 conducting intelligent audits of visa information, and uploading data through intelligent audit algorithms; the audit content focuses on the integrity and authenticity of the visa page, the application form, and the final formation of the file. Unlike the collection stage, the audit of the collection stage focuses on the scope of the file collection and the standardization of the file preparation.

[97] Step 6.3: introducing the image recognition algorithm to extract the signatures involved in the file visa, and standardize the entry of signature data through data preprocessing;

[98] The extracted file visa is divided into two types: offline signature and online signature, the offline signature usually needs to perform operations such as drying, correcting, and smoothing the image; online signature usually requires data point format conversion, standardization of data decimal point accuracy, standardization of sampling frequency, data alignment and other operations.

[99] Step 6.4: extracting global features and local features of the signature data after preprocessing, and comparing them with the features of a registered signature and a verification signature, calculating the feature similarity, and outputting a signature identification result according to a set feature similarity preset value;

[100] Step 6.5: transferring final results of an intelligent audit to an interface of a corresponding audit unit and department. If the audit passes the visa formalities directly through each department, the results are saved to the local database. If it fails to pass, the corresponding reminder information will be sent to the audit unit.

[101] Based on the above method, an intelligent collection system for engineering archives based on enabling thinking, this system includes an intelligent collection module, an intelligent classification module, an intelligent identification module, an intelligent generating test paper module, an intelligent cataloging module, and an intelligent visa module.

[102] Intelligent collection module: it is used to complete the automatic receiving, 7504881 auditing, and registration of other business system interface archive files.

[103] Intelligent classification module: it is used to divide the attribution category and determine the classification number automatically.

[104] Intelligent identification module: it is used to extract intelligent classification results and form rule sets, and then call them through the rule engine, so as to divide the category and determine the classification number automatically.

[105] intelligent generating test paper module: it is used to simulate the process of forming and arranging physical files in reality, and complete intelligent assistance of generating test paper, file sorting, and in-volume file sorting.

[106] Intelligent cataloging module: it is used to realize two independent functions of page number automatic writing and file number automatic generation by integrating deep learning and natural language processing algorithms.

[107] Intelligent visa module: it is used to realize the intelligentization of the visa process of the completion file.

. . . ; . LU504881

[108] Table 1: Engineering project file archiving template accountability

File type Secondary classification unit

Tender invitation 7)

General principles of bidding

Tender Documents En

Bidding Tender Documents documents Tender opening documents En

Bid evaluation documents En

Contract Documents

Technical standards and requirements En owes

Contract agreement En

Notice of winning bid En

Tender Letter and Appendix of Tender

Letter

Contract Special contract terms

Documents General contract terms En

Technical standards and requirements baw

Priced bill of quantities En

Claims

CLAIMS LU504881

1. An intelligent collection method for engineering archives based on enabling thinking, this method includes the following steps: Step 1: intelligent collection, completing an automatic reception, audit, and registration of other business system interface archiving files intelligently through an application of artificial intelligence algorithms; Step 2: intelligent classification, processing, and analyzing text contents of engineering archives with the help of an artificial intelligence algorithm to divide attribution categories and determine classification numbers automatically; Step 3: intelligent identification, extracting intelligent classification results and forming a rule set with the help of an artificial intelligence algorithm, and then calling the results and the rule set through a rule engine to divide attribution categories and determine classification numbers automatically; Step 4: intelligent generating test paper, simulating the process of generating and arranging physical files in reality, completing intelligent assistance of generating test paper, file sorting, and in-volume file sorting; Step 5: intelligent cataloging, realizing two independent functions of page number automatic writing and file number automatic generation by integrating deep learning and natural language processing algorithms; Step 6: intelligent visa, introducing artificial intelligence algorithm to realize the intelligentization of the visa process of the completion file.

2. The intelligent collection method for engineering archives based on enabling thinking according to claim 1, wherein Step 1 also includes the following steps: Step 1.1: adding an automatic receiving program to perform preprocessing; Step 1.2: identifying PDF, ODF format text files, photo files, and other multimedia files by an algorithm, and then using an information extraction algorithm to analyze the morphology and syntax of the identified information;

Step 1.3: entering an intelligent audit link, and eliminating files that do not 7504881 meet the requirements; Step 1.4: starting a registration process, giving an electronic file a unique identifier automatically, using a natural language processing algorithm to extract the form entry, and using a feature matching algorithm to match the file content and the information entry to realize the automatic filling of the information of a registration form.

3. The intelligent collection method for engineering archives based on enabling thinking according to claim 1, wherein Step 2 also includes the following steps: Step 2.1: applying a recognition algorithm to identify registered engineering project documents and multimedia electronic archives; Step 2.2: using a natural language processing algorithm to analyze a basic analysis, an attribute analysis, a semantic analysis, and a structural analysis of an identified engineering archives, finding a relationship model between a file text attribute and a file archiving category according to a pre-set archiving template; Step 2.3: sampling scanned parts in an existing engineering archives collection, cleaning the data according to input requirements of a learning algorithm, constructing a training set and a testing set, and achieving an expected learning effect by constantly improving the model; Step 2.4: inputting registered project files into a relational model, matching the files to a filing scope in a filing module, and further refining a filing category to realize the intelligent classification of a file ownership category in the whole file; Step 2.5: on the basis of accurate classification of file categories, matching a classification number set by an archiving template according to the second- level category name, and supporting an authorized user to set the classification code for the newly created files.

4. The intelligent collection method for engineering archives based on 7504881 enabling thinking according to claim 1, wherein Step 3 also includes the following steps: Step 3.1: creating a rule engine and determine an interface to invoke a rule set; Step 3.2: compiling and loading external business rules into a rule set for the rule engine to call and execute; Step 3.3: extracting results of intelligent classification as a set of data objects processed by rule sets and adding the set to a working memory; Step 3.4: executing a rule matching command by an engine, matching an application object in the working memory with a condition part of the rule in a rule set container, returning a corresponding value of a classification number in an archive template and a storage period table, and deriving a preliminary judgment result.

5. The intelligent collection method for engineering archives based on enabling thinking according to claim 1, wherein Step 4 also includes the following steps: Step 4.1: inputting identified files into a machine learning model; Step 4.2: incorporating set rules of file arrangement and generating test papers in s volume, operation norms, and experience summary in the practice of arrangement and generating test paper into the rule base of deep learning to support the deep learning model; Step 4.3: training and modeling the system by using a file set of sorted and grouped papers to form a computer-readable sorting and a grouped paper rule; Step 4.4: identifying the nature of the file through a sorting law of the files in the volume, and clustering according to the same characteristics to assist generating test papers; extracting topic keywords automatically through a deep learning algorithm, and generating a file title and the key information of the file in combination with title generation rules provided;

Step 4.5: sorting files and files in the volume according to file sorting rules, 7504881 and generating the management of the files and files in the volume automatically.

6. The intelligent collection method for engineering archives based on enabling thinking according to any of claims 1-5, wherein Step 6 also includes the following steps: Step 6.1: uploading archives that need to be reviewed by personnel of all parties according to the requirements of visa processing; Step 6.2: receiving visa information by the system automatically, conducting intelligent audits of visa information, and uploading data through intelligent audit algorithms; Step 6.3: introducing the image recognition algorithm to extract the signatures involved in the file visa, and standardize the entry of signature data through data preprocessing; Step 6.4: extracting global features and local features of the signature data after preprocessing, and comparing them with the features of a registered signature and a verification signature, calculating the feature similarity, and outputting a signature identification result according to a set feature similarity preset value; Step 6.5: transferring final results of an intelligent audit to an interface of a corresponding audit unit and department.

7. The intelligent collection method for engineering archives based on enabling thinking according to claim 2, wherein in step 1.3, the intelligent audit link also includes a determination of the filing scope, a quality audit, and a batch check.

8. An intelligent collection system for engineering archives based on 7504881 enabling thinking, this system includes an intelligent collection module, an intelligent classification module, an intelligent identification module, an intelligent generating test paper module, an intelligent cataloging module, and an intelligent visa module.

9. A computer readable storage medium that stores computer programs, the method according to any of the claims 1 to 7 is implemented when the computer program is executed by a processor.