File classification management method

Abstract

The invention relates to the technical field of archive storage and management, and discloses an archive classification management method which comprises the following steps: constructing an alliance chain property block chain network and an I PFS file system; for the received multi-dimensional electronic archive set, performing clustering analysis on the electronic archive set to obtain an archive classification result of the electronic archive set; based on the number of the electronic archives in the archive classification result, the storage nodes are classified to obtain node classification results, and the node classification results are in one-to-one correspondence with the archive classification results; according to a node classification result, uploading and storing the classified electronic archive file through the corresponding storage node, generating a hash ID table related to the storage node, and classifying the hash ID table; and the block chain network generates the Hash ID table into a block according to the category and a block chain network operation mechanism, and stores the block into the block chain network.

Description

Technical Field

[0001] The present invention relates to the technical field of archive storage and management, and in particular to an archive classification management method. Background Art

[0002] Large organizations, such as corporations, businesses, governments, hospitals (networks), and educational institutions, often use software applications to perform operations, execute transactions, and record many other events that occur in the course of their daily business. This software is often referred to as "enterprise software." Generally, these software packages have evolved over the years to include a large number of enterprise resources to assist with forecasting, reporting, and database management. The comprehensive platform on which these enterprises can fully execute, plan, and investigate problems within the enterprise is called an "Enterprise Resource Planning" system or software, or "ERP" for short.

[0003] Using these enterprise resource planning systems, entities within an organization can complete enterprise-related tasks, and the resulting actions can be recorded on non-transitory computer-readable media for later access and processing or reference. At the organizational scale, the resulting data accumulated over time may be processed in large quantities to generate customized, standard or automated reports, business forecasts, or otherwise make observations and test hypotheses related to the organization. This data is typically stored locally, with off-site automatic backup, or may be stored in the "cloud". For example, a small organization may operate a single enterprise software computer workstation, saving all data to storage media within the computer workstation or directly connected to the computer workstation. Larger organizations may implement more complex storage systems, which in turn become more complex as more agents of the organization access, retrieve, modify and / or otherwise "touch" the organizational file system. The data on these systems can be stored in one or more formats, and currently electronic archive files on "enterprise resource planning" systems or software generally require large-scale human effort to organize, which is inefficient. Summary of the invention

[0004] In view of the above-mentioned prior art, the present invention provides a file classification management method, which mainly solves the technical problems existing in the above-mentioned background technology.

[0005] To achieve the above object, the technical solution of the embodiment of the present invention is implemented as follows:

[0006] A method for file classification management, the method comprising the following steps:

[0007] Constructing a blockchain network with consortium chain properties and an IPFS file system, wherein the blockchain network stores label information of electronic archives, and the IPFS file system has multiple storage nodes for storing the main body information of electronic archives;

[0008] For the received multi-dimensional electronic archive collection, cluster analysis is performed on the electronic archive collection to obtain an archive classification result of the electronic archive collection;

[0009] Based on the number of electronic archives in the archive classification results, the storage nodes are classified to obtain node classification results, and the node classification results correspond to the archive classification results one by one;

[0010] According to the node classification results, the classified electronic archive files are uploaded and stored through the corresponding storage nodes, and a hash ID table about the storage nodes is generated and the hash ID table is classified;

[0011] The blockchain network generates blocks according to the categories of the above hash ID table and saves them to the blockchain network according to the blockchain network operation mechanism.

[0012] Optionally, cluster analysis is performed on the electronic archive collection, specifically including:

[0013] Sequentially extract the name information and content information of any electronic archive file in the electronic archive set, and determine the relevance between the name information and content information of the electronic archive file;

[0014] For electronic archive files with high correlation between name information and content information, cluster analysis is performed based on the name information.

[0015] Optionally, for electronic archive files whose name information and content information have low correlation, all subtitles in the electronic archive are extracted, the title information is input into the Bi LSTM+CNN combined model, and the subtitles are classified according to semantic features to obtain new name information.

[0016] Optionally, determining the correlation between the name information and the content information of the electronic archive file specifically includes converting the name information into a first numerical vector, converting the content information into a second numerical vector, and calculating the cosine distance or Euclidean distance between the first numerical vector and the second numerical vector; if the cosine distance or the Euclidean distance is greater than a threshold, the correlation between the name information and the content information is high; otherwise, the correlation between the name information and the content information is low.

[0017] Optionally, for electronic archive files whose name information and content information are highly correlated, cluster analysis is performed based on the name information, specifically including: using the TF-IDF model to convert the name information into vector information, executing the K-means-I SODATA algorithm on the vector information, and adaptively generating multiple clusters, each cluster representing a classification result.

[0018] Optionally, storage nodes are classified based on the number of electronic archives in the archive classification results to obtain node classification results, specifically including dividing the storage nodes into a corresponding number of categories based on the number of archive classification results, and allocating a certain number of storage nodes to the storage node category based on the number of electronic archives in each archive classification result.

[0019] Optionally, in the process of executing the K-means-I SODATA algorithm on the vector information to adaptively generate multiple clusters, the elbow rule is used to determine the initial K value.

[0020] Optionally, the hash ID table generated by the IPFS file system is the only way for the above block data to be referenced and identified in the IPFS file system. Each block data has a separate link structure. Different block data can be quickly authenticated between different nodes through the link structure. The blockchain network obtains complete archive data from the IPFS file system through the hash ID table.

[0021] Optionally, the hash ID table is classified, specifically including: classifying the hash ID table according to the result of archive classification.

[0022] The beneficial effect of the present invention is that the archive classification management method provided by the present invention combines blockchain technology with the IPFS distributed storage system to realize the automatic classification storage and management of electronic archives. First, by analyzing the name information and content information of the electronic archives, the Bi LSTM+CNN model is used to ensure that the archives can be accurately classified even when the correlation between the name information and the content information is low. Secondly, through the K-means-ISODATA algorithm, multiple clustering clusters can be adaptively generated to improve the accuracy of archive classification. Furthermore, the storage nodes are classified based on the archive classification results, and the allocation of storage nodes is optimized, making the storage more reasonable and efficient. The method implemented by the present invention not only improves the efficiency and speed of archive archiving, but also effectively improves the retrieval and access speed of archives, greatly improves the user experience, reduces storage costs, and also ensures the security and reliability of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a flowchart of the file classification management method in the embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solution of the present invention is further elaborated in detail below in conjunction with the drawings and specific embodiments of the specification. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. In the following description, the expression "some embodiments" is related to a subset of all possible embodiments, but it should be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0025] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features well known in the art are not described.

[0026] It should be understood that the present invention can be implemented in different forms and should not be interpreted as being limited to the embodiments proposed herein. On the contrary, providing these embodiments will make the disclosure thorough and complete, and the scope of the present invention will be fully conveyed to those skilled in the art. And the purpose of the terms used herein is only to describe specific embodiments and is not intended to be a limitation of the present invention. When used herein, the singular forms of "one", "one" and "said / the" are also intended to include plural forms, unless the context clearly indicates another way. It should also be understood that the terms "compose" and / or "include" when used in this specification determine the presence of the features, integers, steps, operations, elements and / or parts, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts and / or groups. When used herein, the term "and / or" includes any and all combinations of the relevant listed items.

[0027] It should also be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "inside", "outside", "left", "right" and similar expressions used herein are for illustrative purposes only and are not intended to be the only implementation method.

[0028] In order to fully understand the present invention, a detailed structure will be proposed in the following description to illustrate the technical solution proposed by the present invention. The optional embodiments of the present invention are described in detail as follows, but in addition to these detailed descriptions, the present invention may also have other implementations.

[0029] Please refer to the attached Figure 1 The present application provides a method for file classification management, the method comprising the following steps:

[0030] S1. Construct a blockchain network with consortium chain properties and an IPFS file system. The blockchain network stores the label information of electronic archives. The IPFS file system has multiple storage nodes for storing the main information of electronic archives. This determines the infrastructure of the entire archive management system and ensures the secure storage and fast access of data.

[0031] S2. For the received multi-dimensional electronic archive collection, cluster analysis is performed on the electronic archive collection to obtain archive classification results of the electronic archive collection;

[0032] S3. Based on the number of electronic archives in the archive classification result, the storage nodes are classified to obtain a node classification result, wherein the node classification result corresponds to the archive classification result one by one;

[0033] S4. According to the node classification result, the classified electronic archive files are uploaded and stored through the corresponding storage nodes, and a hash ID table about the storage nodes is generated and the hash ID table is classified;

[0034] S5. The blockchain network generates blocks based on the categories of the above hash ID table and saves them to the blockchain network according to the blockchain network operation mechanism.

[0035] Specifically, the important factors that affect the speed of file downloads are the communication bandwidth between nodes and the number of storage nodes. In the current situation where communication bandwidth is generally in excess, the number of storage nodes has become the main factor limiting the speed of file downloads. The storage size of a certain data file is 200Mb, and the number of storage nodes in the IPFS file system is 2. Other entities that need to view and download the file can only transmit data from these two nodes. Therefore, the embodiment of the present invention adaptively classifies the categories of electronic archive files, and allocates a corresponding number of storage nodes to the corresponding categories based on the classification results, thereby optimizing the number of nodes, thereby greatly improving the efficiency and speed of file archiving, and effectively improving the speed of downloading and viewing the file.

[0036] In the specific implementation, cluster analysis is performed on the received multi-dimensional electronic archive collection to obtain the archive classification results. The cluster analysis process first involves extracting the name information and content information of each electronic archive file, and judging the correlation between the two. For electronic archive files with high correlation, cluster analysis is performed directly based on the name information. For files with low correlation, all subtitles in the document are extracted, reclassified according to the semantic features of the subtitles, and new name information is obtained. In this way, even if the original file name does not match its content, the system can automatically correct it and classify it correctly;

[0037] After obtaining the classification results of the electronic archive collection, the next step is to classify the storage nodes according to the number of electronic archives in the archive classification results. This process allocates a corresponding number of storage nodes according to the number of electronic archives in each category to ensure that each category of archives can receive efficient storage services. In this way, the system can dynamically adjust storage resources to meet the storage needs of different categories of archives.

[0038] Subsequently, according to the classification results of the storage nodes, the classified electronic archive files are uploaded and stored through the corresponding storage nodes, and a hash ID table about the storage nodes is generated. These hash ID tables are then classified for subsequent management and retrieval.

[0039] Exemplarily, electronic archive files include at least the following: financial documents: such as budget reports, financial statements, audit reports, etc.; personnel documents: such as employee files, attendance records, performance evaluations, etc.; project documents: such as project plans, progress reports, project summaries, etc.; market documents: such as market research reports, marketing plans, advertising materials, etc.; legal documents: such as contracts, agreements, legal opinions, etc.; technical documents: such as technical specifications, product manuals, research and development reports, etc.; administration: such as company rules and regulations, meeting minutes, notices and announcements, etc.

[0040] In an optional implementation, cluster analysis is performed on the electronic archive collection, specifically including:

[0041] Sequentially extract the name information and content information of any electronic archive file in the electronic archive set, and determine the relevance between the name information and content information of the electronic archive file;

[0042] For electronic archive files with high correlation between name information and content information, cluster analysis is performed based on the name information.

[0043] Specifically, if the name information and content information of the electronic archive file are highly correlated, it means that the name and content of the electronic archive file match, so cluster analysis can be performed directly based on the name.

[0044] For electronic archive files with low correlation between name information and content information, the processing idea is to regenerate the name information based on the content information.

[0045] Furthermore, for electronic archive files with low correlation between name information and content information, all subtitles in the electronic archive are extracted, the title information is input into the BiLSTM+CNN combined model, and classification is performed according to the semantic features of the subtitles to obtain new name information.

[0046] The idea of ​​the Bi LSTM+CNN combination model is to connect the features output by each neuron of the Bi LSTM into a feature matrix of the sentence, which is used as the input of the convolutional neural network for secondary feature extraction. Specifically, the word vector sequence composed of the words in the sentence is encoded into a two-dimensional matrix by the Bi LSTM model. The encoding process is the process of learning the recursive model, which filters and fuses the word vector features of the sentence into the feature matrix of the sentence. The characteristic of the recursive model is to capture long-distance text semantic features and extract related features for long-distance semantics, while the Bi LSTM model ensures that the forward and backward semantic information of the sentence can be captured. The sentence feature matrix output by the Bi LSTM model contains the long-distance semantic information between the words in the sentence. TextCNN is used to perform convolution operations on the feature matrix of the sentence, and feature screening is performed through pooling, and then the related features of adjacent words in the sentence feature matrix are extracted.

[0047] For example, when the document name is "Attendance Record" and the content is related to the project summary, the word vector sequence composed of all subtitles in the electronic document is input into the BiLSTM+CNN combined model, and according to the classification result, the standard name of "Project Summary" is obtained. At this time, the electronic document file originally named "Attendance Record" is renamed "Project Summary".

[0048] In an optional implementation, determining the correlation between the name information and the content information of an electronic archive file specifically includes converting the name information into a first numerical vector, converting the content information into a second numerical vector, and calculating the cosine distance or Euclidean distance between the first numerical vector and the second numerical vector; if the cosine distance or the Euclidean distance is greater than a threshold value, the correlation between the name information and the content information is high; otherwise, the correlation between the name information and the content information is low.

[0049] Furthermore, for electronic archive files with high correlation between name information and content information, cluster analysis is performed based on the name information, specifically including: using the TF-IDF model to convert the name information into vector information, executing the K-means-I SODATA algorithm on the vector information, and adaptively generating multiple clusters, each cluster representing a classification result.

[0050] Specifically, the name information is first converted into vector information through the TF-IDF model, and then an initial K value is determined using the elbow rule. Using the determined initial K value, the K-means algorithm is run for preliminary clustering, K samples are randomly selected as initial centroids, each title is assigned to the cluster where the nearest centroid is located, and the centroid of each cluster is updated. The above steps are repeated until the centroid no longer changes significantly or the maximum number of iterations is reached;

[0051] The ISODATA algorithm is applied to merge and split cluster pairs. For merging clusters, all cluster pairs are checked. If the distance between two clusters is less than a certain threshold (such as Euclidean distance or cosine similarity), they can be merged.

[0052] For split clustering, the samples in each cluster are checked. If the samples in a cluster are very different (such as the variance within the cluster is greater than a threshold), it can be split into two or more clusters.

[0053] The final clustering results include finance, personnel, projects, marketing, law, technology, administration and other categories, and the corresponding electronic archive files are assigned to the corresponding categories.

[0054] In an optional implementation, storage nodes are classified based on the number of electronic archives in the archive classification results to obtain node classification results, specifically including dividing the storage nodes into a corresponding number of categories based on the number of archive classification results, and allocating a certain number of storage nodes to the storage node category based on the number of electronic archives in each archive classification result.

[0055] Specifically, based on the clustering results of the archives, clustering groups of storage nodes are generated synchronously. For example, categories such as finance, personnel, projects, marketing, law, technology, and administration are generated synchronously, and files in different groups in the archive classification results are assigned corresponding storage nodes to the groups of storage nodes. For example, in the archive file classification, the number of electronic archive files in the finance, personnel, projects, marketing, law, technology, and administration categories are 10,000, 6,000, 8,450, 20,000, 3,412, 12,789, etc., respectively. Therefore, the number of storage nodes in the finance group is 10, the number of storage nodes in the personnel group is 6, the number of storage nodes in the project group is 8, etc.

[0056] In an optional implementation, in the process of executing the K-means-ISODATA algorithm on the vector information to adaptively generate multiple clusters, the elbow rule is used to determine the initial K value, specifically:

[0057] Run the K-means algorithm multiple times, each time using a different value of K (for example, starting with 2 and gradually increasing).

[0058] For each value of K, the sum of squared distances (SSE) of all points to their nearest centroid is calculated.

[0059] Plot the relationship between K value and SSE, and find the point where the rate of SSE decreases significantly slower, that is, the "elbow". The K value corresponding to this point is the initial number of clusters.

[0060] Furthermore, the hash ID table generated by the IPFS file system is the only way for the above block data to be referenced and identified in the IPFS file system. Each block data has a separate link structure. Different block data can be quickly authenticated between different nodes through the link structure. The blockchain network obtains complete archive data from the IPFS file system through the hash ID table.

[0061] Furthermore, the hash ID table is classified, specifically including: classifying the hash ID table according to the result of the archive classification.

[0062] Specifically, when the archive files are divided into several major categories such as finance, personnel, projects, marketing, law, technology, and administration, the hash ID table is also divided into the aforementioned categories. Each category contains the hash ID of the corresponding node, and blocks are generated by category and saved in the blockchain network.

[0063] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. The protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A file classification management method, characterized in that: The method comprises the following steps: Constructing a blockchain network with consortium chain properties and an IPFS file system, wherein the blockchain network stores label information of electronic archives, and the IPFS file system has multiple storage nodes for storing the main body information of electronic archives; For the received multi-dimensional electronic archive collection, cluster analysis is performed on the electronic archive collection to obtain an archive classification result of the electronic archive collection; Based on the number of electronic archives in the archive classification results, the storage nodes are classified to obtain node classification results, and the node classification results correspond to the archive classification results one by one; According to the node classification results, the classified electronic archive files are uploaded and stored through the corresponding storage nodes, and a hash ID table about the storage nodes is generated and the hash ID table is classified; The blockchain network generates blocks according to the categories of the above hash ID table and saves them to the blockchain network according to the blockchain network operation mechanism.

2. A method for file classification management according to claim 1, characterized in that: Performing cluster analysis on the electronic archive collection, specifically including: Sequentially extract the name information and content information of any electronic archive file in the electronic archive set, and determine the relevance between the name information and content information of the electronic archive file; For electronic archive files with high correlation between name information and content information, cluster analysis is performed based on the name information.

3. A file classification management method according to claim 2, characterized in that: For electronic archive files with low correlation between name information and content information, all subtitles in the electronic archive are extracted, and the title information is input into the BiLSTM+CNN combined model, and classified according to the semantic features of the subtitles to obtain new name information.

4. A method for file classification management according to claim 3, characterized in that: Determining the correlation between the name information and the content information of the electronic archive file specifically includes converting the name information into a first numerical vector, converting the content information into a second numerical vector, and calculating the cosine distance or Euclidean distance between the first numerical vector and the second numerical vector. If the cosine distance or the Euclidean distance is greater than a threshold, the correlation between the name information and the content information is high; otherwise, the correlation between the name information and the content information is low.

5. A method for file classification management according to claim 4, characterized in that: For electronic archive files with high correlation between name information and content information, cluster analysis is performed based on the name information, specifically including: using the TF-IDF model to convert the name information into vector information, executing the K-means-ISODATA algorithm on the vector information, and adaptively generating multiple clusters, each cluster representing a classification result.

6. A method for file classification management according to claim 5, characterized in that: Based on the number of electronic archives in the archive classification results, the storage nodes are classified to obtain node classification results, specifically including: based on the number of archive classification results, the storage nodes are divided into a corresponding number of categories, and based on the number of electronic archives in each archive classification result, a certain number of storage nodes are allocated to the storage node category.

7. A method for file classification management according to claim 6, characterized in that: In the process of executing the K-means-ISODATA algorithm on the vector information to adaptively generate multiple clusters, the elbow rule is used to determine the initial K value.

8. The file classification management method according to claim 1, characterized in that: The hash ID table generated by the IPFS file system is the only way for the above block data to be referenced and identified in the IPFS file system. Each block data has a separate link structure. Different block data can be quickly authenticated between different nodes through the link structure. The blockchain network obtains complete archive data from the IPFS file system through the hash ID table.

9. The file classification management method according to claim 8, characterized in that: Classifying the hash ID table specifically includes: classifying the hash ID table according to the result of the archive classification.

Images