CN118069791B

CN118069791B - Intelligent electronic archive retrieval method and system

Info

Publication number: CN118069791B
Application number: CN202410482666.9A
Authority: CN
Inventors: 赵显瑞; 陈东升; 于观杰; 任伟彬; 李燕云; 贾秋月; 董庆友
Original assignee: Heze Social Insurance Service Center; Heze Product Inspection And Testing Institute
Current assignee: Heze Social Insurance Service Center; Heze Product Inspection And Testing Institute
Priority date: 2024-04-22
Filing date: 2024-04-22
Publication date: 2024-07-05
Anticipated expiration: 2044-04-22
Also published as: CN118069791A

Abstract

The invention relates to the technical field of data retrieval, in particular to an intelligent electronic archive retrieval method and system, wherein the method comprises the following steps: acquiring real-time electronic archive data and user retrieval demand data; carrying out semantic feature analysis on the real-time electronic archive data to generate electronic archive semantic feature data; performing word frequency distribution analysis on the semantic feature data of the electronic file to generate word frequency distribution data; performing word frequency vectorization on the word frequency distribution data to generate word frequency vector data; space mapping is carried out on the semantic feature data of the electronic archive through word frequency vector data, and archive vector points are generated; carrying out multidimensional space reconstruction on the archive vector points to construct an electronic archive multidimensional space; and carrying out vector point dynamic association analysis on the multidimensional space of the electronic file to generate a vector point dynamic association network. The invention realizes efficient and accurate electronic archive retrieval.

Description

Intelligent electronic archive retrieval method and system

Technical Field

The invention relates to the technical field of data retrieval, in particular to an intelligent electronic archive retrieval method and system.

Background

Along with the rapid development of information technology and the wide application of electronic files, the traditional manual search method often has the problems of low search efficiency and poor accuracy facing massive electronic file data, and can not meet the requirements of users on efficient and accurate search, so that an electronic file intelligent search method and an electronic file intelligent search system based on natural language processing and machine learning are generated.

Disclosure of Invention

The invention provides an intelligent electronic file retrieval method and system for solving at least one technical problem.

In order to achieve the above object, the present invention provides an intelligent electronic archive retrieval method, comprising the following steps:

Step S1: acquiring real-time electronic archive data and user retrieval demand data; carrying out semantic feature analysis on the real-time electronic archive data to generate electronic archive semantic feature data; performing word frequency distribution analysis on the semantic feature data of the electronic file to generate word frequency distribution data;

Step S2: performing word frequency vectorization on the word frequency distribution data to generate word frequency vector data; space mapping is carried out on the semantic feature data of the electronic archive through word frequency vector data, and archive vector points are generated; carrying out multidimensional space reconstruction on the archive vector points to construct an electronic archive multidimensional space;

Step S3: carrying out vector point dynamic association analysis on the multidimensional space of the electronic file to generate a vector point dynamic association network; performing time sequence morphological change analysis on the multidimensional space of the electronic file to generate space time sequence morphological change data; carrying out evolution trend analysis on the space time sequence morphological change data according to the vector point dynamic association network so as to obtain a multidimensional space time-space evolution rule;

Step S4: carrying out multi-layer index space construction on the electronic archive multidimensional space by utilizing a multidimensional space-time evolution rule to construct a dynamic multidimensional space-time archive space; performing space-time characteristic point conversion on the user retrieval demand data to generate retrieval demand space-time characteristic points; carrying out search range analysis on the dynamic multidimensional space-time archives by utilizing the space-time characteristic points of the search requirements to generate search range data;

step S5: carrying out retrieval space boundary extraction on the dynamic multidimensional space-time archive space by utilizing retrieval range data to obtain an archive retrieval space; performing search key feature analysis on the user search demand data to generate search key feature data; performing optimal retrieval processing on the file retrieval space according to the retrieval key characteristic data to generate file space optimal retrieval data; carrying out iterative retrieval on the dynamic multidimensional space-time archive space according to the archive space optimal retrieval data so as to generate iterative retrieval space path data;

Step S6: performing evolution track analysis on the iterative search space path data to generate search evolution track data; carrying out search path optimization decision on the dynamic multidimensional space-time archive space by utilizing search evolution track data to generate search path optimization decision data; and constructing a graph neural network for the search path optimization decision data, and constructing a archive search knowledge graph so as to execute the intelligent search operation of the electronic archive.

The invention can deeply understand the semantic meaning of the electronic archive data and the importance of the key words through semantic feature analysis and word frequency distribution analysis, the generated electronic archive semantic feature data and word frequency distribution data provide a basis for subsequent retrieval and analysis, word frequency vectorization converts the word frequency distribution data into a vector form, the subsequent vector calculation and similarity comparison are facilitated, the electronic archive semantic feature data is converted into vector points in a multidimensional space through space mapping and multidimensional space reconstruction, a basis is provided for subsequent association analysis and space-time evolution, the vector point dynamic association analysis and the generated dynamic association network can reveal the association relation and interaction in the multidimensional space of the electronic archive, the spatial time sequence morphological change analysis and evolution trend analysis help to know the change and trend of the multidimensional space of the electronic archive along with time, the space-time characteristics of the archive data are disclosed, the multi-layer index space construction is convenient for efficient searching and inquiring operation, the searching speed and accuracy are improved, the conversion of the space-time characteristic points of the searching requirement and the generation of the searching range data are helpful for determining the searching range and space, the searching range is reduced, the searching efficiency is improved, the searching space boundary extraction is used for determining the searching range and boundary of the archive by utilizing the searching range data, thus unnecessary calculation and traversal are reduced, the searching key characteristic analysis and the generation of the searching key characteristic data are helpful for determining the key points and the attention points of the searching requirement of a user, the searching accuracy and the relevance are improved, the optimal searching processing of the archive space and the generation of the iterative searching space path data are improved by optimizing the searching algorithm and path selection, the generation and analysis of the search evolution track data can reveal the evolution trend and the association relation in the search process, is beneficial to understanding the search behavior and improving the search strategy, optimizes the search path and the strategy by utilizing the search evolution track data, improves the search efficiency and accuracy, integrates the association relation of the search path optimization decision data and the archive data by utilizing the graph neural network in the construction of the archive search knowledge graph, and provides deeper search support and an intelligent archive search function.

In this specification, an electronic archive intelligent search system is provided, configured to perform the electronic archive intelligent search method described above, including:

The word frequency distribution module is used for acquiring real-time electronic archive data and user retrieval demand data; carrying out semantic feature analysis on the real-time electronic archive data to generate electronic archive semantic feature data; performing word frequency distribution analysis on the semantic feature data of the electronic file to generate word frequency distribution data;

The multidimensional space module is used for carrying out word frequency vectorization on the word frequency distribution data so as to generate word frequency vector data; space mapping is carried out on the semantic feature data of the electronic archive through word frequency vector data, and archive vector points are generated; carrying out multidimensional space reconstruction on the archive vector points to construct an electronic archive multidimensional space;

The evolution trend module is used for carrying out vector point dynamic association analysis on the multidimensional space of the electronic file so as to generate a vector point dynamic association network; performing time sequence morphological change analysis on the multidimensional space of the electronic file to generate space time sequence morphological change data; carrying out evolution trend analysis on the space time sequence morphological change data according to the vector point dynamic association network so as to obtain a multidimensional space time-space evolution rule;

The search range module is used for constructing a multi-layer index space for the multi-dimensional space of the electronic file by utilizing a multi-dimensional space time evolution rule to construct a dynamic multi-dimensional space-time file space; performing space-time characteristic point conversion on the user retrieval demand data to generate retrieval demand space-time characteristic points; carrying out search range analysis on the dynamic multidimensional space-time archives by utilizing the space-time characteristic points of the search requirements to generate search range data;

The retrieval space path module is used for extracting retrieval space boundaries of the dynamic multidimensional space-time archive space by utilizing the retrieval range data to obtain an archive retrieval space; performing search key feature analysis on the user search demand data to generate search key feature data; performing optimal retrieval processing on the file retrieval space according to the retrieval key characteristic data to generate file space optimal retrieval data; carrying out iterative retrieval on the dynamic multidimensional space-time archive space according to the archive space optimal retrieval data so as to generate iterative retrieval space path data;

The knowledge graph module is used for carrying out evolution track analysis on the iterative search space path data so as to generate search evolution track data; carrying out search path optimization decision on the dynamic multidimensional space-time archive space by utilizing search evolution track data to generate search path optimization decision data; and constructing a graph neural network for the search path optimization decision data, and constructing a archive search knowledge graph so as to execute the intelligent search operation of the electronic archive.

The invention obtains real-time electronic archive data and user retrieval demand data through a word frequency distribution module, provides a data basis for the subsequent steps, performs semantic feature analysis on the real-time electronic archive data, extracts semantic features of the archive data, such as keywords, topics and the like, performs word frequency distribution analysis on the electronic archive semantic feature data, counts the occurrence frequency of each word in the archive data, generates word frequency distribution data, performs word frequency vectorization on the word frequency distribution data by a multidimensional space module, converts the word frequency of each word into vector representation, generates word frequency vector data, performs space mapping on the electronic archive semantic feature data through the word frequency vector data, maps the semantic feature data into multidimensional space, generates archive vector points, performs multidimensional space reconstruction on the archive vector points, reorganizes and represents the archive vector points in the multidimensional space, constructing a multidimensional space structure of the electronic file, carrying out vector point dynamic association analysis on the multidimensional space of the electronic file by an evolution trend module, analyzing association relations among different vector points, generating a vector point dynamic association network, carrying out time sequence morphological change analysis on the multidimensional space of the electronic file, researching the time sequence morphological change rule of the space, generating space time sequence morphological change data, carrying out evolution trend analysis on the space time sequence morphological change data based on the vector point dynamic association network, revealing the time-space evolution rule of the multidimensional space, such as change trend, periodicity and the like, carrying out multi-layer index space construction on the multidimensional space of the electronic file by a search range module by utilizing the multidimensional space time-space evolution rule, constructing a dynamic multidimensional space-time file space, providing a high-efficiency search range, carrying out time-space feature point conversion on user search requirement data, converting the search demand of a user into space-time feature points, representing the search demand position of the user in a multi-dimensional space, carrying out search range analysis on a dynamic multi-dimensional space-time archive space by utilizing the search demand space-time feature points, determining an archive range matched with the user demand, generating search range data, carrying out search space boundary extraction on the dynamic multi-dimensional space-time archive space by utilizing the search range data, determining the boundary of the archive search space, reducing the search range, improving the search efficiency, carrying out search key feature analysis on the user search demand data, extracting key features such as key words and attributes of the user search demand, carrying out optimal search processing on the archive search space according to the search key feature data, determining an optimal search path, generating archive space optimal search data, carrying out iterative search on the dynamic multi-dimensional space-time archive space according to the archive space optimal search data, progressively searching archive data according to the optimal path, generating iterative search space path data, carrying out evolution path analysis on the iterative search space path data, researching the change trend and law of the search path data, carrying out path optimization decision making decision on the dynamic multi-dimensional space archive space, carrying out path optimization search on the search decision making decision, generating a search path optimization decision graph according to path optimization decision graph, and carrying out construction of a neural network, and carrying out neural network optimization search map by utilizing the neural network.

Drawings

FIG. 1 is a flowchart illustrating steps of an intelligent electronic archive retrieval method according to the present invention;

FIG. 2 is a detailed implementation step flow diagram of step S1;

FIG. 3 is a detailed implementation step flow diagram of step S2;

Fig. 4 is a detailed implementation step flow diagram of step S3.

Detailed Description

It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

The embodiment of the application provides an intelligent electronic file retrieval method and system. The execution main body of the electronic archive intelligent retrieval method and the system comprises, but is not limited to, the system: mechanical devices, data processing platforms, cloud server nodes, network uploading devices, etc. may be considered general purpose computing nodes of the present application, including but not limited to: at least one of an audio image management system, an information management system and a cloud data management system.

Referring to fig. 1 to 4, the present invention provides an intelligent electronic file searching method, which includes the following steps:

In the embodiment of the present invention, as described with reference to fig. 1, a flow chart of steps of an intelligent electronic archive retrieval method of the present invention is shown, and in this example, the steps of the intelligent electronic archive retrieval method include:

in this embodiment, real-time electronic archive data and user retrieval demand data are acquired; when a user submits a search request to a system, related data are obtained, which can be keywords, query sentences, filtering conditions and the like input by the user, accurate information required by the user is ensured to be obtained, semantic feature analysis is carried out on the obtained real-time electronic archive data, semantic information in a text can be extracted by using Natural Language Processing (NLP) technology, such as parts of speech tagging, entity recognition, keyword extraction and the like, the semantic feature of the archive data can be analyzed, information of topics, keywords, attributes and the like related to the archive can be recognized, the result obtained by the semantic feature analysis is converted into structured data, the electronic archive semantic feature data can be generated, which can be a data structure stored in the forms of text, JSON, XML and the like, the keyword and the feature of the archive are included, the occurrence frequency of each word in the archive data is counted, the occurrence frequency of each word can be preprocessed by using text processing technology, such as word segmentation, stop word removal and the like, the occurrence frequency of each word is counted, and the archive frequency distribution data is formed, and the importance degree of each word in the data is indicated.

In this embodiment, the word Frequency distribution data is converted into word Frequency vector data, the word Frequency vector is a way of representing the word Frequency distribution by using a vector, a word Bag model (Bag-of-Words) or TF-IDF (Term Frequency-Inverse Document Frequency) technology can be used to convert the Frequency of each word in the word Frequency distribution data into a vector form, the word Frequency vector can be a dense vector or a sparse vector, depending on the dimension of the corresponding vocabulary, the word Frequency vector data is utilized to spatially map the electronic archive semantic feature data, the archive semantic feature is mapped into a multidimensional vector space, multidimensional space reconstruction is performed on the mapped archive vector points to construct a multidimensional space of the electronic archive, a multidimensional technology (such as principal component analysis and t-SNE) can be used to convert a high-dimensional vector into a vector representation with lower dimension, so that visualization and analysis can help find similarity and relevance between archives, and more possibilities are provided for subsequent archive retrieval and analysis.

In this embodiment, a dynamic association analysis is performed on vector points in a multidimensional space of an electronic file, the association between points is explored, which can be implemented by calculating indexes such as similarity, distance or correlation between vector points, and based on the result of the association analysis, a dynamic association network of vector points is constructed, nodes in the network represent vector points, edges represent the association relationship between points, weights represent the association degree, and further analysis and visualization can be performed on the dynamic association network by using graph theory and network analysis methods, such as community discovery, centrality analysis, etc., a time-series morphological change analysis is performed on the multidimensional space of the electronic file, evolution and change trend of the archive in time are explored, which can be the change of position, density or other features of the archive in the multidimensional space, the time sequence analysis method, the clustering algorithm or the nonlinear dynamics model and the like can be used for modeling and analyzing the time sequence morphological change of the archive to generate space time sequence morphological change data, the time sequence change information of the archive in a multidimensional space is recorded, the time sequence morphological change data can be a data set taking time as a dimension, each time point corresponds to the characteristic or the state of the archive, the evolution trend analysis is carried out on the space time sequence morphological change data based on a vector point dynamic association network, the time-space evolution rule of the archive can be researched through the change of nodes in the network, the evolution of association edges and the like, and the analysis such as pattern recognition, trend prediction or anomaly detection can be carried out on the vector point dynamic association network and the time-space evolution data by using the methods such as graph theory, time sequence analysis, machine learning or deep learning.

In this embodiment, multi-layer index space construction is performed on the multi-dimensional space of the electronic file based on the space-time evolution rule of the multi-dimensional space, which may be to divide the multi-dimensional space into different levels or regions so as to perform retrieval and management more efficiently, construct a dynamic multi-dimensional space-time file space, integrate the space-time information of the electronic file into the multi-dimensional space, which may be to add a time dimension or other space-time features at each point of the multi-dimensional space so as to reflect the space-time attribute of the file, perform space-time feature point conversion on the user retrieval demand data, map the retrieval demand into space-time feature points, which may be to convert the retrieval condition, time range or other space-time elements of the user into a specific vector representation, use a similar vectorization method, such as word frequency vectorization or feature extraction technology, convert the retrieval demand data into a vector form, perform retrieval range analysis on the dynamic space-time feature point by using the retrieval demand space-time feature points, which may be to determine a space region or range related to the retrieval demand in the multi-dimensional space-time space, analyze the space-time feature point dynamic space, which may be based on a vector point dynamic association network, an evolution method or a space-time-space-time feature point, or a space-time feature region related region, which may be related to the retrieval demand data may be generated, and a space-time feature region, which may be related to the retrieval requirement space-space demand.

in this embodiment, the search space boundary extraction is performed on the dynamic multidimensional space-time archive space by using the search range data generated in the previous step, which may be determining a boundary or limit of the archive space to define a further search range, after the search space boundary is extracted, the archive search space is obtained, which is a defined dynamic multidimensional space-time archive space related to a search requirement, the search key feature analysis is performed on the user search requirement data to determine the key feature of the user search, which may be extracting a keyword, a time range, a spatial region or other features related to the search requirement, generating the search key feature data, recording the key feature of the user search requirement, which may be representing the key feature of the user search requirement in a specific data structure, vector representation or other forms, performing the optimal search processing on the archive search space according to the search key feature data, which may be using an index structure, a query optimization algorithm or a machine learning method to improve the search efficiency and accuracy, performing the optimal search processing, which may be recording the space position, feature or other search result information of the archive space related to the optimal search, and iteratively recording the search result information in the new path, which may be the current path, and performing the iterative search process, which may be the iterative search process, or the new search path-related to the file space-time-space, and the current path-space-related to the iterative search process.

In this embodiment, the evolution track analysis is performed on the iterative search space path data, which may be to analyze patterns, trends or other features in the path data, so as to understand the evolution of the search process, generate search evolution track data, record the evolution track in the search process, which may be to represent the evolution of the search path in a sequence, a time sequence or other forms, perform a search path optimization decision on a dynamic multidimensional space-time archive space by using the search evolution track data, which may be to determine how to optimize the search path to improve the search efficiency or accuracy according to the analysis result of the evolution track, generate search path optimization decision data, record path optimization decision data for different search situations, which may be to represent the path optimization decision data in a rule, a strategy, a parameter configuration or other forms, and perform a graph neural network construction by using the search path optimization decision data, which may be to construct a graph neural network model with learning capability by using the path optimization decision data as node features or side weights, construct a archive search knowledge graph, combine the graph neural network with other related data structures or knowledge bases, which may be to optimize the search path information in the search process by using the graph, such as to form a graph, and provide an intelligent search result, which may be based on the information in the search result, the search process, the map, and the intelligent search process, which may be performed by using the graph search result, and the intelligent search result.

In this embodiment, as described with reference to fig. 2, a detailed implementation step flow diagram of the step S1 is described, and in this embodiment, the detailed implementation step of the step S1 includes:

Step S11: acquiring real-time electronic archive data and user retrieval demand data;

Step S12: performing metadata structuring processing on the real-time electronic archive data to obtain electronic archive structured data;

step S13: carrying out semantic feature analysis on the structured data of the electronic file to generate semantic feature data of the electronic file;

Step S14: performing word frequency calculation on the semantic feature data of the electronic file to generate semantic feature word frequency data;

Step S15: and performing word frequency distribution analysis on the semantic feature word frequency data to generate word frequency distribution data.

The invention can ensure that the used data is up-to-date by acquiring the real-time electronic archive data, reflect the current archive information, acquire the user retrieval demand data, know the requirements and demands of users so as to carry out corresponding retrieval and provide accurate results, convert the electronic archive data into a form with a certain structure, facilitate the subsequent analysis and processing, provide the basic framework and the attribute of the archive data for the generated electronic archive structured data, provide the basis for the subsequent semantic analysis and retrieval, deeply understand the semantic meaning and the key features of the electronic archive data by semantic feature analysis, provide the extraction and the representation of the semantic information of the archive data for the subsequent retrieval and analysis, count the occurrence frequency of semantic feature words in the electronic archive by word frequency calculation, reflect the importance degree of the semantic feature words in the archive, provide the importance degree ordering of different semantic feature words in the archive, provide the basis for the subsequent word frequency distribution analysis and the retrieval weighting, and provide the basis for the semantic feature words in the electronic archive, and the semantic feature word frequency analysis can provide the importance degree ordering of the semantic feature words in the electronic archive, and the relevant feature distribution and the information.

In this embodiment, the real-time electronic archive data may be obtained from an electronic archive database, a file system or other data sources, so as to ensure that the data obtaining process has real-time performance, so as to obtain the latest electronic archive data, obtain the retrieval requirement data of the user through a user interface, an API interface or other channels, ensure that the obtained user retrieval requirement data contains necessary information, such as keywords, a time range, a space region and the like, perform structural processing on the extracted metadata, organize the extracted metadata into a data structure capable of being queried and analyzed, map the metadata into a table, a relational database or other data structures, perform semantic feature extraction on the electronic archive structural data, capture semantic information of text, use Natural Language Processing (NLP) technology, extracting features by using methods such as word vectors and topic models to generate semantic feature data of an electronic archive, organizing the extracted semantic features into a data format which can be used for subsequent analysis, wherein the semantic features can be expressed as vectors, matrixes or other forms of data structures, performing word frequency calculation on each feature in the semantic feature data of the electronic archive, counting the occurrence frequency of each feature in a document, calculating the word frequency in a simple counting mode, generating semantic feature word frequency data by using a more complex weighted calculation method, organizing the statistical result of the feature word frequency into a data structure which can be used for subsequent analysis, the feature word frequency can be expressed as vectors, matrixes or other forms of data structures, performing word frequency distribution analysis on the semantic feature word frequency data, knowing the distribution situation of the word frequency in an integral data set, and calculating the average word frequency, statistical indexes such as standard deviation of word frequency, maximum word frequency, minimum word frequency and the like are used for generating word frequency distribution data, and the result of word frequency distribution analysis is organized into a data format which can be used for subsequent analysis and visualization, wherein the word frequency distribution can be represented as a histogram, a box diagram, a probability density diagram or other forms of data structures.

In this embodiment, as described with reference to fig. 3, a detailed implementation step flow diagram of the step S2 is shown, and in this embodiment, the detailed implementation step of the step S2 includes:

Step S21: performing word frequency vectorization on the word frequency distribution data to generate word frequency vector data;

Step S22: space mapping is carried out on the semantic feature data of the electronic archive through word frequency vector data, and archive vector points are generated;

step S23: performing spatial position analysis on the archive vector points to generate archive vector point spatial position data;

Step S24: similarity calculation is carried out on the archive vector points so as to generate vector point similarity data;

Step S25: clustering and dividing the archive vector points through the vector point similarity data to generate archive clustering data;

step S26: and carrying out multidimensional space reconstruction on the archive cluster data according to the archive vector point space position data so as to construct an electronic archive multidimensional space.

The invention converts word frequency distribution data into a vector form through word frequency vectorization, the word frequency of each semantic feature word is expressed as the dimension of a vector, the generated word frequency vector data provides the position information of each semantic feature word in a vector space, a basis is provided for subsequent vector calculation and similarity analysis, the electronic archive semantic feature data is converted into a vector point form through space mapping, the semantic feature of each archive is expressed as a point in the vector space, the generated archive vector point provides the position information of archive data in the vector space, a basis is provided for subsequent similarity calculation and cluster analysis, the spatial position analysis can know the distribution condition of archive vector points in the vector space, reflect the relative position relation among different archives, the generated archive vector point spatial position data provides the coordinate information of the archive vector points, providing a basis for subsequent similarity calculation and cluster analysis, wherein the similarity calculation can measure the similarity degree between archive vector points and is used for determining the relevance and similarity between archives, the generated vector point similarity data provides similarity evaluation between archive vector points and provides a basis for subsequent clustering and retrieval, the cluster division can group similar archive vector points into different clusters so as to classify similar archives into one class, the generated archive cluster data provides a cluster division result of the archives and provides a basis for subsequent cluster analysis and intelligent retrieval, the multidimensional space reconstruction can visually display the archive cluster data in a multidimensional space so as to facilitate the user to understand and browse electronic archive data, the constructed electronic archive multidimensional space provides multidimensional display and navigation, and visual archive browsing and intelligent retrieval experience is provided for users.

In this embodiment, the word frequency distribution data is converted into word frequency vectors, the word frequency of each word is used as one dimension of the vector,

The method of using vectorization technology, such as Bag-of-Words (TF-IDF) and TF-IDF, can be used to transform the word Frequency distribution into vector representation, generate word Frequency vector data, organize the word Frequency vector into a data structure that can be used for subsequent analysis, which can be represented as vector matrix or other form of data structure, space map the electronic file semantic feature data by using the word Frequency vector data, represent the semantic feature of each document as a vector point, the mapping can use the method of dimension reduction technology such as Principal Component Analysis (PCA), t-SNE (t-Distributed Stochastic Neighbor Embedding) and the like to reduce dimension and preserve the structure of the semantic feature, generate file vector points, organize the space mapping result of each document into a data structure that can be used for subsequent analysis, which can be represented as vector matrix or other form of data structure, analyze the space position of the file vector points, understand the distribution condition of the vector points in multidimensional space, calculate the coordinate of each vector point, distance, calculate the position of the spatial position point, and the like, calculate the position of the spatial position point, and the position of the spatial point can be used as a measure the similarity between the spatial point and the spatial point, the position of the file, the similarity is calculated by the method of the spatial point, the similarity is similar to the spatial point of the spatial point, the spatial position of the spatial point is calculated, the spatial point is similar to the spatial point is calculated, and the spatial position of the spatial point is calculated, and the spatial position is used for the spatial position of the subsequent analysis is calculated, the result of similarity calculation is organized into a data structure which can be used for subsequent analysis and query, the similarity can be expressed as a similarity matrix, a similarity graph or other forms of data structures, the vector points are clustered by using vector point similarity data, the similar vector points are grouped into clusters, clustering can be carried out by using a clustering algorithm such as K-means, hierarchical clustering (HIERARCHICAL CLUSTERING) and the like, multidimensional space reconstruction can be carried out on the archive cluster data according to archive vector point space position data, the archive cluster can be visually displayed in a multidimensional space by using a visual technology such as a scatter diagram, a thermodynamic diagram, a 3D visual and the like, and the multidimensional space representation of the electronic archive is constructed according to the result of the multidimensional space reconstruction, which can be expressed as an interactive visual interface, a multidimensional space index structure or other forms of data structures, so that a user can browse, query and analyze the electronic archive.

In this embodiment, as described with reference to fig. 4, a detailed implementation step flow diagram of the step S3 is shown, and in this embodiment, the detailed implementation step of the step S3 includes:

step S31: carrying out vector point dynamic association analysis on the multidimensional space of the electronic file to construct a vector point dynamic association network;

Step S32: performing time sequence analysis on the multidimensional space of the electronic file to generate multidimensional space time sequence data;

Step S33: performing time sequence form change analysis on the multidimensional space time sequence data to generate space time sequence form change data;

Step S34: performing space-time evolution simulation on the space time sequence morphological change data according to the vector point dynamic association network to generate multidimensional space evolution simulation data;

step S35: and carrying out evolution trend analysis on the multidimensional space evolution simulation data so as to obtain a multidimensional space time-space evolution rule.

According to the invention, through vector point dynamic association analysis, association relations among vector points in the archive data can be revealed, the similarity, the correlation and the like are included, the constructed vector point dynamic association network can describe association network structures among different vector points in the archive data, a basis is provided for subsequent retrieval and recommendation, time sequence analysis can know variation trends and evolution conditions of the archive data at different time points, generated multidimensional space time sequence data provides variation information of the archive data in time dimension, basis is provided for subsequent time sequence form variation analysis and evolution simulation, time sequence form variation analysis can reveal form variation and trend of the archive data in multidimensional space, generated spatial time sequence form variation data provides form evolution information of the archive data in multidimensional space, basis is provided for subsequent time-space evolution simulation and trend analysis, time-space evolution simulation can simulate evolution processes of the archive data in multidimensional space, the generated multidimensional space simulation data provides evolution processes of the archive data in time-space dimension, basis is provided for subsequent evolution analysis and intelligent retrieval, evolution trend analysis can reveal form variation and trend in multidimensional space, the generated spatial time-space evolution form variation data can be provided for prediction rules of the archive data in the multidimensional space, and the development trend is provided for the prediction of the relevant evolution rules and the like.

In this embodiment, a dynamic association analysis is performed on vector points in a multidimensional space of an electronic archive, association relationships and evolution trends among the vector points are explored, the association degree among the vector points can be calculated by using methods such as correlation analysis, time sequence analysis and the like, a vector point dynamic association network is constructed based on the dynamic association of the vector points, the vector points can be represented as nodes of the network, the dynamic association can be represented as edges among the nodes, the time sequence analysis is performed on the vector points in the multidimensional space of the electronic archive, evolution changes of the vector points in time can be observed, time sequence features of the vector points such as mean values, variances, trends and the like can be calculated, multidimensional space time sequence data is generated, the time sequence analysis results are organized into data structures which can be used for subsequent analysis and visualization, the multidimensional space time sequence data can be represented as time sequences, data tables or other forms of data structures, performing time sequence shape change analysis on multi-dimensional space time sequence data, exploring shape evolution of the data on time sequence, extracting shape characteristics of the time sequence data by using methods such as shape analysis, pattern recognition and the like, generating space time sequence shape change data, organizing results of the shape change analysis into a data structure which can be used for subsequent analysis and visualization, wherein the time sequence shape change data can be represented as a shape characteristic sequence, a shape change graph or other forms of data structures, performing space time evolution simulation on the space time sequence shape change data according to a vector point dynamic association network, simulating the evolution process of vector points in the multi-dimensional space, simulating the space time evolution relationship among the vector points by using methods such as a network model, a simulation algorithm and the like, generating multi-dimensional space evolution simulation data, the result of the space-time evolution simulation is organized into a data structure which can be used for subsequent analysis and visualization, the evolution simulation data can be expressed into a time sequence, a multidimensional space evolution graph or other forms of data structures, the evolution trend analysis is carried out on the multidimensional space evolution simulation data, the evolution rule and trend of the data are explored, the evolution trend can be analyzed by using methods such as statistical analysis and machine learning, and the like, the law of multidimensional space-time evolution can be obtained based on the result of the evolution trend analysis, and the laws can be related to the relation among vector points, the evolution trend, periodicity and other aspects.

In this embodiment, the specific steps of step S31 are as follows:

step S311: performing content association analysis on the multidimensional space of the electronic archive to generate multidimensional space content association data;

Step S312: carrying out association trend analysis on the multidimensional space content association data to generate association strong and weak trend data;

step S313: performing association transfer identification on the association strong and weak trend data to generate vector point association transfer data;

Step S314: carrying out space span analysis on the multidimensional space of the electronic archive by using vector point association transfer data so as to generate vector point space span data;

step S315: carrying out vector point dynamic association analysis on the vector point space span data to generate vector point dynamic association data;

Step S316: and carrying out association network fitting on the vector point dynamic association data to construct a vector point dynamic association network.

The invention can reveal the association relation between the contents of the archive data in the multidimensional space through content association analysis, comprising semantic similarity, theme correlation and the like, the generated multidimensional space content association data provides association information of the archive data on the content dimension, provides basis for subsequent association trend analysis and association transfer identification, the association trend analysis can know the strong and weak variation trend of the content association in the archive data, the generated association strong and weak trend data provides dynamic variation information of the content association in the archive data, provides basis for subsequent association transfer identification and space span analysis, the association transfer identification can discover the transfer mode and trend of the content association in the archive data, the generated vector point association transfer data provides basis for subsequent space span analysis and dynamic association analysis, the space span analysis can reveal the span and distribution condition of the vector point in the archive data on the space, the generated vector point space analysis provides basis for subsequent vector point dynamic association analysis and association network fitting, the vector point association analysis can be provided with the subsequent vector point association point in the archive data, the generated vector point association information can be provided for the information in the archive data, the association relation in the network has the basis of the information of the association relation of the content association in the archive data can be provided, the association relation in the network has the information is provided, the association relation of the information in the network is provided by the association relation of the information in the space, the method provides a basis for intelligent retrieval of the electronic files, and comprises the functions of associated recommendation, similar file discovery and the like.

In this embodiment, association analysis is performed on contents in a multidimensional space of an electronic archive, association relations between the contents are explored, association relation of text data can be extracted and analyzed by using methods such as text mining, natural language processing and the like, multidimensional space content association data is generated, the result of the content association analysis is organized into a data structure which can be used for subsequent analysis and processing, the content association data can be represented as an association matrix, an association graph or other forms of data structure, association trend analysis is performed on the multidimensional space content association data, association trend can be observed, association trend data can be analyzed by using methods such as statistical analysis, time series analysis and the like, association trend data is generated, the result of the association trend analysis is converted into a data structure which can be used for subsequent analysis and processing, the association trend data can be represented as a time series, a trend graph or other forms of data structure, association transition identification is performed on the association trend data, a transfer mode and a machine learning method can be used for identifying association transition, vector point association transition data can be generated, the association trend point association trend is identified as a transfer point can be used for subsequent analysis, the association point can be used for processing the association graph, the association graph can be used for the space span of the data, the largest-span of the space span can be used for the space span analysis, the data can be used for calculating the association graph, the association graph can be used for the space span, the space span of the data can be used for processing, and the data can be used for processing the space span, and the data can be obtained by using the space span, and the space span of the data can be used for the space analysis, the method can be used for representing the space span data as a vector point span sequence, a span graph or other forms of data structures, carrying out vector point dynamic association analysis on the vector point space span data, exploring the dynamic association relation between vector points, and calculating the dynamic association degree between the vector points by using methods such as correlation analysis, time sequence analysis and the like.

In this embodiment, the specific steps of step S35 are as follows:

Step S351: performing time activation curve analysis on the multidimensional space evolution simulation data to generate a multidimensional space time activation curve;

Step S352: carrying out structure complexity evolution analysis on the multidimensional space of the electronic file according to the multidimensional space time activation curve so as to generate multidimensional space structure complexity evolution data;

Step S353: edge effect detection is carried out on the multi-dimensional space structure complexity evolution data to generate edge structure effect data;

step S354: performing implicit association analysis on the multidimensional space of the electronic archive based on the edge structure effect data to generate multidimensional space implicit association data;

Step S355: carrying out regional time sequence change analysis on the file cluster data based on the multidimensional space implicit correlation data so as to generate cluster regional time sequence change data;

step S356: and carrying out evolution trend analysis on the multidimensional space evolution simulation data based on the time sequence change data of the cluster area, thereby obtaining a multidimensional space time evolution rule.

The invention can reveal the change trend of the evolution of the archival data along with time in the multidimensional space through time activation curve analysis, the generated multidimensional space time activation curve provides activation information of the archival data along with time dimension, can help understand the time sequence relationship and evolution rule of the archival data, the structure complexity evolution analysis can evaluate the change condition of the structure complexity of the archival data along with time in the multidimensional space, the generated multidimensional space structure complexity evolution data provides quantitative information of the archival data structure evolution, can help understand the structure change and complexity trend of the archival data, the edge effect detection can reveal the influence degree of the edge region in the archival data on the whole structure, the generated edge structure effect data provides importance information of the edge region in the archival data, can help identify and understand key structure edges in the archival data, the implicit association analysis can find potential association relations in the archive data, even if the association relations are not obvious in the original data, the generated multidimensional space implicit association data provides the implicit association information in the archive data, can help reveal potential association and association modes between the archive data, the regional time sequence change analysis can explore the evolution and change trend of cluster areas in the archive data along with time, the generated cluster area time sequence change data provides the time-space evolution information of the cluster areas in the archive data, can help understand the time sequence characteristics and change rules of the clusters in the archive data, the evolution trend analysis can summarize the time-space evolution rules and trends of the multidimensional space in the archive data, the evolution trend of the archive data on time-space can be obtained through the analysis of the cluster area time sequence change data, important insights are provided about the spatiotemporal characteristics of archival data.

In this embodiment, the time activation curve analysis is performed on the multidimensional space evolution simulation data to observe the changes of activity levels of different dimensions along with time, the time activation curve of the multidimensional space can be calculated and drawn by using methods such as time sequence analysis, statistical analysis and the like, the multidimensional space time activation curve is generated, the result of the time activation curve analysis is organized into a data structure which can be used for subsequent analysis and processing, the time activation curve can be represented as time sequence data, a graph or other forms of data structures, the structure complexity evolution analysis is performed on the multidimensional space of the electronic archive based on the multidimensional space time activation curve, the complexity index of the spatial structure is calculated by using methods such as complex network analysis and graph theory to study the complexity change of the spatial structure, the edge effect detection is performed on the complexity evolution data of the multidimensional space structure, the edge region in the spatial structure can be identified by using methods such as edge detection algorithm and cluster analysis, the hidden association analysis is performed on the multidimensional space of the electronic archive based on the edge structure effect data, the potential association relationship can be found, the hidden association relationship can be used for processing the hidden association relationship in the time sequence analysis data, the hidden association relationship can be generated in the time sequence analysis and the hidden association data can be used for processing the time sequence analysis, the hidden association data can be clustered or the hidden association data can be generated by using the hidden association relationship in the time sequence analysis, the time sequence change of the cluster area is analyzed by methods such as cluster change detection, the evolution trend analysis is carried out on the multidimensional space evolution simulation data based on the time sequence change data of the cluster area, the space time evolution rule is revealed, the evolution trend and rule can be analyzed by methods such as trend analysis and pattern recognition, the space time evolution rule of the multidimensional space is obtained according to the result of the evolution trend analysis, the time change mode and trend of the space are described, and the space time evolution rule can be expressed as a rule, a model or other description.

In this embodiment, the specific steps of step S4 are as follows:

Step S41: carrying out multi-layer index space construction on the electronic archive multidimensional space by utilizing a multidimensional space-time evolution rule to construct a dynamic multidimensional space-time archive space;

Step S42: performing space-time characteristic point conversion on the user retrieval demand data to generate retrieval demand space-time characteristic points;

Step S43: carrying out fuzzy space retrieval on the dynamic multidimensional space-time archive space by utilizing the retrieval demand space-time characteristic points so as to generate fuzzy space retrieval data;

Step S44: carrying out semantic feature space distribution analysis on the dynamic multidimensional space-time archive space to generate semantic feature space distribution data;

Step S45: and carrying out search range analysis on the semantic feature space distribution data based on the fuzzy space search data to generate search range data.

The invention utilizes the multi-dimensional space time evolution rule to organize and index the electronic archive data according to different dimensionalities and time, improves the retrieval efficiency of the data, the constructed dynamic multi-dimensional space time archive space can adapt to the evolution and change of the archive data, provides more flexible and dynamic data storage and organization modes for intelligent retrieval, the space-time feature point conversion converts the retrieval requirement data of a user from characters or other forms into feature points with space-time attributes, can better express the space-time retrieval requirement of the user, the generated retrieval requirement space-time feature points provide space-time representation for the retrieval requirement of the user, provides input data for subsequent fuzzy space retrieval, utilizes the retrieval requirement space-time feature points and the index structure of the dynamic multi-dimensional space-time archive space, retrieving data related to user demands from a archive space through fuzzy matching and similarity measurement, providing a archive data set matched with the user demands by the generated fuzzy space retrieval data, providing a basis for further analysis and presentation, analyzing the semantic feature space distribution analysis by carrying out semantic analysis and feature extraction on the data in the dynamic multidimensional space-time archive space, analyzing the distribution condition of the data on the semantic feature space, revealing the semantic relevance of the archive data, providing the distribution information of the archive data on the semantic feature space by the generated semantic feature space distribution data, providing a basis for subsequent retrieval range analysis, analyzing the distribution range of the archive data on the semantic feature space according to the fuzzy space retrieval result and the semantic feature space distribution data, determining the data range meeting the retrieval demands, the generated search range data provides range information of the archive data meeting the search requirements of users, can be used for filtering and displaying results, and improves the accuracy and efficiency of search.

In this embodiment, the multi-layer index space is constructed for the multi-dimensional space of the electronic archive by utilizing the multi-dimensional space time evolution rule obtained before, a space segmentation algorithm (such as a quadtree and an octree) or a grid division method can be used for constructing the multi-layer index space, a dynamic multi-dimensional space time archive space is constructed based on the multi-layer index space, the electronic archive data is mapped to a corresponding space region, the data can be distributed to different space regions according to the space-time attribute of the archive to form a dynamic space-time archive space, the retrieval requirement data of the user is obtained, the data describes the requirement of the user on the space-time feature points of the archive, the space-time feature points can comprise a time range, a space range, keywords or other attributes interested by the user, the space-time feature point conversion is performed for the retrieval requirement data of the user, the user requirement is converted into a space-time feature point representation which can be used for retrieval, the time range can be converted into a time period, the space range can be converted into space coordinates or space regions, and the like, search requirement space-time characteristic points can be generated according to the space-time characteristic point conversion result and used for subsequent fuzzy space search, the search requirement space-time characteristic points can be expressed as a data structure of the time period, the space coordinates or other forms, the search requirement space-time characteristic points are utilized for fuzzy space search on the dynamic multidimensional space-time archive space to obtain archive data matched with requirements, fuzzy search algorithms (such as fuzzy query and fuzzy match) can be used for realizing fuzzy space search, fuzzy space search data are generated according to the fuzzy space search result, the archive data matched with search requirements are included, and the data can be expressed as an identifier, an attribute value or other forms of data structures of the archive, the method comprises the steps of carrying out semantic feature space distribution analysis on a dynamic multidimensional space-time archive space, researching the distribution situation of different semantic features in the space, analyzing the distribution of the semantic features by using methods such as cluster analysis, thermodynamic diagram analysis and the like, generating semantic feature space distribution data according to the result of the semantic feature space distribution analysis, describing the space distribution situation of different semantic features, wherein the data can be expressed as the association of the semantic features with a space region, the distribution density of the semantic features or other forms of data structures, carrying out search range analysis on the semantic feature space distribution data based on fuzzy space search data, analyzing which semantic features are more concentrated or sparse in search results and the distribution range of the semantic features in the space, generating search range data according to the result of the search range analysis, describing the search range of each semantic feature in the space, and expressing the data as the association of the semantic feature with the space range, the dense region of the semantic feature or other forms of data structures.

In this embodiment, the specific steps of step S5 are as follows:

Step S51: carrying out retrieval space boundary extraction on the dynamic multidimensional space-time archive space by utilizing retrieval range data to obtain an archive retrieval space;

Step S52: performing demand feature analysis on the user retrieval demand data to generate user demand feature data;

step S53: performing search key feature analysis on the user demand feature data to generate search key feature data;

Step S54: performing optimal retrieval processing on the file retrieval space according to the retrieval key characteristic data to generate file space optimal retrieval data;

Step S55: and carrying out iterative retrieval on the dynamic multidimensional space-time archive space according to the archive space optimal retrieval data so as to generate iterative retrieval space path data.

According to the invention, range information of file data meeting the user retrieval requirement is provided through retrieval range data, boundary extraction can be carried out on dynamic multidimensional space-time file space by utilizing the information, boundary extraction of the file retrieval space is beneficial to narrowing the retrieval range, uncorrelated data interference is reduced, retrieval accuracy and efficiency are improved, demand feature analysis can be used for carrying out deep understanding and analysis on the retrieval requirement data of a user, key features in the demand feature analysis can be extracted, generated user demand feature data provides key feature information of the user retrieval requirement, a basis is provided for subsequent retrieval processing, retrieval key feature analysis is used for further analysis on the user demand feature data, the most important feature for file retrieval is extracted, key feature information for optimal retrieval processing is provided for generating retrieval key feature data, accuracy and correlation of retrieval results are improved, optimal retrieval processing is used for screening and sequencing the file retrieval space by utilizing the retrieval key feature data, data most conforming to the user demand is arranged in front, the generated file space optimal retrieval data provides file data set according to the key feature sequencing, retrieval requirement of the user can be better met, the iterative retrieval path is provided for the iterative retrieval path of the more optimal retrieval data by utilizing the optimal retrieval space, and the path is more accurate and the optimal retrieval path is provided for the user retrieval path is more accurate and is more suitable for the user retrieval path-ordered.

In this embodiment, based on the search range data, search space boundaries of the dynamic multidimensional space-time archive space are extracted, where the boundaries may be the minimum and maximum ranges of space coordinates or represent the boundaries of a space region, and according to the result of the search space boundary extraction, an archive search space is obtained, the space defines the space range that needs to be considered in the search process, demand feature analysis is performed on user search demand data, key demand features of users are extracted and analyzed, text analysis, keyword extraction and other techniques can be used to identify important features in user demand, according to the result of demand feature analysis, user demand feature data is generated, the main demand features of users are described, these data may be represented as a keyword list, a data structure of attribute requirements or other forms of data, search key feature analysis is performed on user demand feature data, these features may be represented as the most important attribute, keyword or other identifier when searching an archive, key feature data is generated according to the result of search key feature analysis, these data may be represented as a keyword list, data of required data structure or other forms of data is identified, the matching with the optimal search result of search profile can be achieved by matching the optimal search result, the profile can be processed according to the optimal search result, the matching is achieved, the optimal search result is achieved, and the profile is matched with the optimal search result is obtained, the attribute value or other forms of data structures are used for carrying out iterative search operation based on file space optimal search data to obtain more accurate search results, the file space can be subjected to iterative search according to factors such as user requirements, data characteristics and the like, iterative search space path data are generated according to the iterative search results, paths and sequences for carrying out iterative search in the file space are described, and the data can be expressed as a series of space positions, time steps or other forms of path data.

In this embodiment, the specific steps of step S6 are as follows:

step S61: performing search evolution analysis on the iterative search space path data to generate iterative search evolution data;

Step S62: performing evolution track analysis on the iterative search evolution data to generate search evolution track data;

step S63: carrying out search path optimization decision on the dynamic multidimensional space-time archive space by utilizing search evolution track data to generate search path optimization decision data;

Step S64: and constructing a graph neural network for the search path optimization decision data, and constructing a archive search knowledge graph so as to execute the intelligent search operation of the electronic archive.

According to the invention, the iterative search space path data is subjected to deep research and analysis through search evolution analysis, the evolution rule and the change trend are explored, the generated iterative search evolution data provides evolution information of the path data, the analysis and arrangement of the path change in the iterative search evolution data are facilitated, the data representation of an evolution track is formed through evolution track analysis, the generated search evolution track data provides time sequence information of the path change, the change trend and the evolution mode of the archive data can be revealed, a search path optimization decision is based on the search evolution track data, the decision making of path optimization is carried out through analysis of the evolution trend and the historical data of the path, the generated search path optimization decision data provides an optimized search path, the search efficiency and accuracy can be improved, the more refined search requirement of a user is met, the graph neural network is constructed based on the search path optimization decision data, the data is converted into a graph structure, the training and construction of the graph neural network are carried out, the constructed archive search knowledge graph can store and represent the relation and the characteristics among the archive data, and the deep understanding and the reasoning capability are provided for intelligent search of the electronic archive.

In this embodiment, search evolution analysis is performed on iterative search space path data, evolution modes and trends in the path data are explored, techniques such as data mining and statistical analysis can be used to analyze changes and evolution of the path data, iterative search evolution data is generated according to the result of the search evolution analysis, evolution features and trends of the path data are described, the data can be expressed as evolution modes, statistical indexes of the path changes or other forms of data, evolution track analysis is performed on the iterative search evolution data, tracks and modes in the evolution data are researched, techniques such as data visualization and cluster analysis can be used to explore the tracks and modes in the evolution data, search path optimization decision is performed based on the search evolution track data, and an optimal search path in a dynamic multidimensional space-time archive space is found, path optimization decisions can be performed by using methods such as heuristic algorithms, optimization algorithms and the like, search path optimization decision data is generated according to the result of the search path optimization decisions, the search path after optimization is described, the data can be expressed as a path sequence, a path optimization index or other forms of data, a graph neural network model is constructed based on the search path optimization decision data and used for modeling knowledge and relations in file search tasks, a graph neural network algorithm (such as GCN and GAT) can be used for constructing a model, a file search knowledge graph is constructed according to the graph neural network model and used for representing relations and attributes among file data, the knowledge graph can comprise file data nodes, attribute nodes and relation edges to form a structured graph data representation, the constructed file search knowledge graph is utilized for performing intelligent search tasks of electronic files, the intelligent retrieval function can be realized by using methods such as a graph query algorithm, a knowledge graph reasoning and the like, and file data meeting the conditions can be found by querying nodes and edges in the knowledge graph according to the retrieval requirements of users.

In this embodiment, an electronic archive intelligent search system is provided, which is configured to execute the electronic archive intelligent search method described above, and includes:

The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

It will be understood that, although the terms "first," "second," etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of example embodiments. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.

The foregoing is only a specific embodiment of the invention to enable those skilled in the art to understand or practice the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent retrieval method for electronic files is characterized by comprising the following steps:

2. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S1 are as follows:

3. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S2 are as follows:

4. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S3 are as follows:

5. The intelligent retrieval method for electronic files according to claim 4, wherein the specific steps of step S31 are as follows:

6. The intelligent retrieval method for electronic files according to claim 4, wherein the specific steps of step S35 are as follows:

7. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S4 are as follows:

8. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S5 are as follows:

9. The intelligent retrieval method of electronic files according to claim 1, wherein the specific steps of step S6 are as follows:

10. An electronic archive intelligent retrieval system for executing the electronic archive intelligent retrieval method as claimed in claim 1, comprising: