An electronic file management method, system and electronic equipment for tobacco industry project management

Abstract

The present invention relates to the field of archive management technology, specifically to an electronic archive management method, system and electronic equipment for tobacco industry project management, including: parsing the contents of multiple electronic archives entered into the archive management system to obtain electronic archive texts, and preprocessing the electronic archive texts to obtain standard electronic archive texts; extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text, thereby obtaining the subject corresponding to the electronic archive, and integrating the standard electronic archive text based on the text subject to obtain integrated electronic archive text. The present invention can better extract standard electronic archive texts by parsing and preprocessing the contents of multiple electronic archives entered into the archive management system, thereby improving the quality and availability of text information.

Description

Technical Field

[0001] The present invention relates to the technical field of archive management, and in particular to an electronic archive management method, system and electronic equipment for project management in the tobacco industry. Background Art

[0002] Tobacco projects usually refer to projects related to tobacco production, sales and related fields. These projects may involve various aspects such as tobacco planting, production and processing of tobacco products, marketing, health promotion, etc. In some countries, tobacco projects may also involve issues such as tobacco industry supervision, tax policies and public health; electronic files of tobacco projects usually include various electronic files and data related to tobacco projects, such as information on planting, production, sales, marketing, supervision, research, etc.

[0003] Under traditional methods, the analysis, integration, classification and display of electronic archive texts require a lot of manpower and time investment, and the efficiency is low. In addition, due to the lack of intelligent text subject extraction and integration, users often need to spend a lot of time on information retrieval, which reduces the availability of information and management efficiency. In addition, the archive management system under traditional methods usually lacks intuitive visual display methods, making it difficult for users to intuitively understand and manage archive information. Moreover, traditional methods usually cannot achieve real-time updating and dynamic display of archive information, and it is difficult to reflect the latest archive changes in a timely manner. Summary of the invention

[0004] The purpose of the present invention is to solve the problems existing in the background technology and to propose an electronic archive management method, system and electronic equipment for project management in the tobacco industry.

[0005] The technical solution of the present invention is an electronic file management method for tobacco industry project management, comprising:

[0006] Parsing the contents of multiple electronic archives entered into the archive management system to obtain electronic archive texts, and performing preprocessing operations on the electronic archive texts to obtain standard electronic archive texts;

[0007] Extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text, thereby obtaining the subject corresponding to the electronic archive, and integrating the standard electronic archive text based on the text subject to obtain an integrated electronic archive text;

[0008] Performing text classification on the integrated electronic archive text by using a trained text classification model to obtain a category corresponding to the integrated electronic archive text, thereby obtaining a category corresponding to the electronic archive;

[0009] Constructing a 3D archive bookshelf model based on three-dimensional modeling technology, and constructing a 3D archive model corresponding to the electronic archive, constructing a category label of the 3D archive bookshelf model based on the category corresponding to the electronic archive, and constructing a subject label of the 3D archive model based on the subject corresponding to the electronic archive;

[0010] Integrate the category label with the 3D archive bookshelf model to obtain a 3D label archive bookshelf model, integrate the subject label with the 3D archive model to obtain a 3D label archive model, and merge the 3D label archive model with the 3D label archive bookshelf model to obtain a 3D archive comprehensive model;

[0011] Based on the adding and deleting operations of the plurality of electronic archives in the archive management system, the 3D archive comprehensive model is updated to obtain a 3D dynamic archive comprehensive model, and the 3D dynamic archive comprehensive model is displayed.

[0012] Preferably, extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text includes:

[0013] The number of topics to be generated is determined according to the size and perplexity of the standard electronic archive text. The perplexity index calculation formula is as follows:

[0014]

[0015] Among them, Perplexity (D) represents the perplexity of the standard electronic archive text, p (w) represents the probability of each word appearing in the standard electronic archive text, N d represents the total number of words in the dth paragraph of the standard electronic archive text, d represents the dth paragraph of the standard electronic archive text, M represents the total number of paragraphs in the standard electronic archive text, and p(w)=p(z|d)*p(w|z), p(z|d) represents the probability of the dth paragraph of the standard electronic archive text being displayed under the text theme, and p(w|z) represents the probability of a word being displayed under the text theme;

[0016] Extracting the topic-word multinomial distribution parameters of the words of the text topic in the standard electronic archive text respectively according to Dirichlet distribution;

[0017] Extracting the text topic distribution in the standard electronic archive text according to Dirichlet distribution based on topic-word multinomial distribution parameters to obtain the final text topic distribution and text topic word distribution;

[0018] By judging the topic similarity of the word distribution in the text topic of each of the standard electronic archive texts, the topics with similar topics are merged to obtain the text topic corresponding to the standard electronic archive text. The topic similarity calculation formula is as follows:

[0019]

[0020] Among them, sim(T m ,T n ) represents the first text topic T m and the second text topic T n The similarity between mi and p ni Respectively represent the first text topic T m and the second text topic T n The probability distribution value of the i-th dimension topic word, v represents the first text topic T m and the second text topic T n The dimension of the words to be compared in the vector space.

[0021] Preferably, the standard electronic archive text is integrated based on the text theme to obtain an integrated electronic archive text, including

[0022] Creating an abstract tag table, wherein the abstract tag table includes a plurality of abstract tags, and establishing a mapping relationship table between the abstract tag table and the text topics;

[0023] According to the mapping relationship table, the standard electronic archive text is mapped and matched with the abstract label table to obtain the abstract label corresponding to the standard electronic archive text;

[0024] The abstract tag is matched with a preset tag content library to obtain the tag content corresponding to the abstract tag, and the tag content is added to the corresponding standard electronic archive text to obtain the integrated electronic archive text.

[0025] Preferably, the text classification model includes a text representation layer, a feature extraction layer and a fully connected layer. The text representation layer adopts the ERNIE model. The feature extraction layer adopts the BiGRU model to extract global feature information, and introduces an attention mechanism into the BiGRU model. The CNNs model is used to extract local feature information, and the global feature information and the local feature information are fused to obtain fused feature information. The fully connected layer performs dimensionality reduction processing on the fused feature information, and predicts the category of the integrated electronic archive text through a classifier.

[0026] Preferably, the BiGRU model includes a forward GRU and a reverse GRU, and the forward hidden layer state calculation formula of the forward GRU at time t is: in, represents the forward hidden layer state and the reverse hidden layer state at time t-1, x t represents the input at time t, f GRU (·) represents the hidden layer calculation function of the unidirectional GRU, and the reverse hidden layer state calculation formula of the reverse GRU at time t is: in, represents the reverse hidden layer state at time t-1, and the attention mechanism calculation formula is: Among them, H t represents the feature information of the forward GRU and reverse GRU of the fusion attention mechanism at time t, f att (·) represents the calculation function of the attention mechanism, and the classifier calculation formula is y=softmax(W con F+b con ), where y represents the probability that the integrated electronic archive text belongs to all categories, softmax(·) represents the activation function, and W con represents the weight matrix, F represents the fusion feature information, b con represents the bias matrix.

[0027] Preferably, the merging of the 3D label archive model and the 3D label archive bookshelf model to obtain a 3D archive comprehensive model comprises:

[0028] storing all the 3D tag archive models in a list to be merged, and sorting them according to the subject tags of the 3D tag archive models;

[0029] Determine whether the 3D label archive model already exists in the 3D label archive bookshelf model space;

[0030] If it does not exist in the 3D label file bookshelf model space, create a strip space whose height is the same as the height of the 3D label file model, and put the 3D label file model into the strip space;

[0031] If it exists in the 3D label archive bookshelf model space, create a horizontal space, and sequentially take out the 3D label archive models from the to-be-merged list and put them into the horizontal space;

[0032] The strip space is placed into the 3D label archive bookshelf model space, thereby obtaining the 3D archive comprehensive model.

[0033] The technical solution of the present invention is an electronic archive management system for tobacco industry project management, which is applicable to the above-mentioned electronic archive management method for tobacco industry project management, including:

[0034] The archive parsing module is used to import multiple electronic archives into the archive management system, parse their contents to obtain electronic archive texts, and perform preprocessing operations to obtain standard electronic archive texts;

[0035] A subject extraction module, the subject extraction module is used to extract the subject of the standard electronic archive text to obtain the text subject, and integrate the standard electronic archive text based on the text subject to obtain the integrated electronic archive text;

[0036] A text classification module, which is used to classify the integrated electronic archive text to obtain the category corresponding to the electronic archive;

[0037] A three-dimensional modeling module, which is used to construct a 3D archive bookshelf model and a 3D archive model corresponding to the electronic archive, and to construct category labels and subject labels;

[0038] A label integration module, the label integration module is used to integrate the category label with the 3D archive bookshelf model to obtain the 3D label archive bookshelf model, integrate the subject label with the 3D archive model to obtain the 3D label archive model, and merge the 3D label archive bookshelf model and the 3D label archive model to obtain the 3D archive comprehensive model;

[0039] The dynamic display module is used to update the 3D archive comprehensive model according to the addition and deletion operations of the electronic archives in the archive management system, obtain the 3D dynamic archive comprehensive model, and display it.

[0040] The technical solution of the present invention: an electronic device, which is applicable to the electronic archive management method for tobacco industry project management, comprises: at least one processor and a memory;

[0041] The memory is used to store computer-executable instructions;

[0042] The at least one processor is used to execute the computer-executable instructions stored in the memory to implement the above-mentioned electronic archive management method for tobacco industry project management.

[0043] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects:

[0044] 1. The present invention can better extract standard electronic archive texts by parsing and preprocessing the contents of multiple electronic archives entered into the archive management system, thereby improving the quality and availability of text information. By extracting according to the subject of the standard electronic archive text and integrating the text subject, it is helpful to classify and organize the electronic archives, thereby improving the efficiency of information retrieval and management. By matching abstract tags with a preset tag content library, the content of the electronic archive can be made richer and more accurate, thereby improving the comprehensibility and accuracy of the information.

[0045] 2. The present invention constructs a 3D archive bookshelf model based on three-dimensional modeling technology, constructs a 3D archive model corresponding to the electronic archive, and constructs corresponding labels for categories and topics, so as to make the archive information more intuitive, easy to understand and manage. By updating and displaying the 3D archive comprehensive model according to the addition and deletion operations of multiple electronic archives in the archive management system, the real-time and dynamic nature of the archive information is maintained. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A schematic flow chart of an overall method in an embodiment of the present invention;

[0047] Figure 2 A schematic diagram of a process of extracting a topic in an embodiment of the present invention;

[0048] Figure 3 A flowchart of text integration in an embodiment of the present invention;

[0049] Figure 4 A schematic diagram of a flow chart of model merging in an embodiment of the present invention;

[0050] Figure 5 A schematic diagram of the overall system flow in an embodiment of the present invention;

[0051] Figure 6 The present invention is a flowchart of an electronic device in an embodiment of the present invention.

[0052] Figure numerals: 1. Archive parsing module; 2. Topic extraction module; 3. Text classification module; 4. Three-dimensional modeling module; 5. Label integration module; 6. Dynamic display module; 7. Memory; 8. Processor; 9. Bus. DETAILED DESCRIPTION

[0053] Embodiment 1, as Figure 1 As shown, the present invention proposes an electronic archive management method for tobacco industry project management, comprising:

[0054] S1. parsing the contents of multiple electronic archives entered into the archive management system to obtain electronic archive texts, and performing preprocessing operations on the electronic archive texts to obtain standard electronic archive texts;

[0055] S2. extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text, thereby obtaining the subject corresponding to the electronic archive, and integrating the standard electronic archive text based on the text subject to obtain the integrated electronic archive text;

[0056] S3, classifying the integrated electronic archive text by using the trained text classification model to obtain the category corresponding to the integrated electronic archive text, thereby obtaining the category corresponding to the electronic archive;

[0057] S4, constructing a 3D archive bookshelf model based on three-dimensional modeling technology, and constructing a 3D archive model corresponding to the electronic archive, constructing a category label of the 3D archive bookshelf model based on the category corresponding to the electronic archive, and constructing a subject label of the 3D archive model based on the subject corresponding to the electronic archive;

[0058] S5, integrating the category label and the 3D archive bookshelf model to obtain a 3D label archive bookshelf model, integrating the subject label with the 3D archive model to obtain a 3D label archive model, and merging the 3D label archive model and the 3D label archive bookshelf model to obtain a 3D archive comprehensive model;

[0059] S6. Based on the adding and deleting operations of multiple electronic archives in the archive management system, the 3D archive comprehensive model is updated to obtain a 3D dynamic archive comprehensive model, and the 3D dynamic archive comprehensive model is displayed.

[0060] In the present invention, the archive management system is a software system for managing and organizing files, records and materials, which helps users effectively store, retrieve, share and protect various types of documents and information; preprocessing operations on electronic archive texts refer to a series of standardization and cleaning operations on text data before performing text analysis, mining or other natural language processing tasks.

[0061] Embodiment 2, as Figure 2-4 As shown, the electronic archive management method for tobacco industry project management proposed by the present invention, compared with the first embodiment, this embodiment also includes: extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text, including:

[0062] A1. Determine the number of topics to be generated based on the size and perplexity of the standard electronic archive text. The perplexity index calculation formula is as follows:

[0063]

[0064] Among them, Perplexity (D) represents the perplexity of the standard electronic archive text, p (w) represents the probability of each word appearing in the standard electronic archive text, N d represents the total number of words in the dth paragraph of the standard electronic archive text, d represents the dth paragraph of the standard electronic archive text, M represents the total number of paragraphs in the standard electronic archive text, and p(w)=p(z|d)*p(w|z), p(z|d) represents the probability of the dth paragraph of the standard electronic archive text being displayed under the text theme, and p(w|z) represents the probability of a word being displayed under the text theme;

[0065] A2, extracting the topic-word multinomial distribution parameters of the words in the standard electronic archive text according to Dirichlet distribution;

[0066] A3. Extract the text topic distribution in the standard electronic archive text according to the Dirichlet distribution based on the topic-word multinomial distribution parameters to obtain the final text topic distribution and the word distribution within the text topic;

[0067] A4. By judging the topic similarity of the word distribution in the text topics of each standard electronic archive text, the topics with similar topics are merged to obtain the text topic corresponding to the standard electronic archive text. The topic similarity calculation formula is as follows:

[0068]

[0069] Among them, sim(T m ,T n ) represents the first text topic T m and the second text topic T n The similarity between mi and p ni Respectively represent the first text topic T m and the second text topic T n The probability distribution value of the i-th dimension topic word, v represents the first text topic T m and the second text topic T n The dimension of the words to be compared in the vector space.

[0070] In this embodiment, perplexity is an indicator used to evaluate the prediction ability of a language model in natural language processing. In the above description, the perplexity indicator is used to determine the number of generated topics. Dirichlet distribution is a multivariate probability distribution, which is often used to describe the probability distribution of multiple random variables. In text topic modeling, Dirichlet distribution is usually used to describe the relationship between text topics and words. The topic-word multinomial distribution parameter is used to describe the probability distribution of words under each topic, that is, the probability of each word appearing in each topic.

[0071] In an optional embodiment, the standard electronic archive text is integrated based on the text theme to obtain an integrated electronic archive text, including

[0072] B1. Create an abstract tag table, which includes multiple abstract tags, and establish a mapping relationship table between the abstract tag table and text topics;

[0073] B2. Map and match the standard electronic archive text with the abstract label table according to the mapping relationship table to obtain the abstract label corresponding to the standard electronic archive text;

[0074] B3. Match the abstract tag with the preset tag content library to obtain the tag content corresponding to the abstract tag, and add the tag content to the corresponding standard electronic archive text to obtain the integrated electronic archive text.

[0075] It should be noted that abstract tags are usually obtained by summarizing and refining the text theme, which are used to represent the theme or content involved in the text. In the text integration process, abstract tags can help classify and organize the text; the mapping relationship table is used to record the correspondence between the abstract tag table and the text theme. Through the mapping relationship table, the text theme can be mapped to the corresponding abstract tag; the tag content library contains the specific content or information corresponding to various tags. After obtaining the abstract tag, it can be mapped to the corresponding tag content by matching the tag content library, thereby enriching the text information.

[0076] In an optional embodiment, the text classification model includes a text representation layer, a feature extraction layer and a fully connected layer. The text representation layer adopts the ERNIE model, the feature extraction layer adopts the BiGRU model to extract global feature information, and introduces the attention mechanism in the BiGRU model. The CNNs model is used to extract local feature information, and the global feature information and the local feature information are fused to obtain fused feature information. The fully connected layer performs dimensionality reduction processing on the fused feature information, and uses a classifier to predict the category of the integrated electronic archive text.

[0077] It should be noted that BiGRU is the abbreviation of Bidirectional Gated Recurrent Unit, which is a commonly used recurrent neural network model. Compared with traditional recurrent neural networks, BiGRU combines forward and backward information to better capture global features in sequence data. In text processing, the BiGRU model is often used to extract global information from text sequences to help understand the semantics and contextual relationships of texts. CNNs stands for Convolutional Neural Networks, which is a commonly used neural network architecture mainly used in the field of image processing. In text processing, CNNs models are also widely used to extract local feature information of texts. Through convolution and pooling operations, CNNs models can effectively capture local patterns and structures in texts, which helps to identify important features in texts.

[0078] In an optional embodiment, the BiGRU model includes a forward GRU and a reverse GRU, and the forward hidden layer state calculation formula of the forward GRU at time t is: in, represents the forward hidden layer state and the reverse hidden layer state at time t-1, x t represents the input at time t, f GRU (·) represents the hidden layer calculation function of the unidirectional GRU. The reverse hidden layer state calculation formula of the reverse GRU at time t is in, represents the reverse hidden layer state at time t-1, and the attention mechanism calculation formula is Among them, H t represents the feature information of the forward GRU and reverse GRU of the fusion attention mechanism at time t, f att (·) represents the calculation function of the attention mechanism, and the classifier calculation formula is y = softmax(W con F+b con ), where y represents the probability that the integrated electronic archive text belongs to all categories, softmax(·) represents the activation function, and W con represents the weight matrix, F represents the fusion feature information, b con represents the bias matrix.

[0079] In an optional embodiment, the 3D label archive model and the 3D label archive bookshelf model are merged to obtain a 3D archive comprehensive model, including:

[0080] C1. Save all 3D tag archive models into the list to be merged, and sort them according to the subject tags of the 3D tag archive models;

[0081] C2, determine whether the 3D tag archive model already exists in the 3D tag archive bookshelf model space;

[0082] C3. If the 3D label archive bookshelf model does not exist in the space, a strip space is created, the height of the strip space is the same as the height of the 3D label archive model, and the 3D label archive model is placed in the strip space;

[0083] C4. If the 3D label archive model exists in the bookshelf model space, create a horizontal space, and take out the 3D label archive model from the list to be merged and put it into the horizontal space in turn;

[0084] C5. Place the strip space into the 3D label archive bookshelf model space to obtain a 3D archive comprehensive model.

[0085] Embodiment three, as Figure 5 As shown, the electronic archive management system for tobacco industry project management proposed by the present invention is applicable to an electronic archive management method for tobacco industry project management, including:

[0086] The archive parsing module 1 is used to import multiple electronic archives into the archive management system, parse their contents to obtain electronic archive texts, and perform preprocessing operations to obtain standard electronic archive texts;

[0087] Theme extraction module 2 is used to extract the theme of the standard electronic archive text to obtain the text theme, and integrate the standard electronic archive text based on the text theme to obtain the integrated electronic archive text;

[0088] The text classification module 3 is used to classify the text of the integrated electronic archive to obtain the category corresponding to the electronic archive;

[0089] The three-dimensional modeling module 4 is used to construct a 3D archive bookshelf model and a 3D archive model corresponding to the electronic archive, and to construct category labels and subject labels;

[0090] The label integration module 5 is used to integrate the category label with the 3D archive bookshelf model to obtain the 3D label archive bookshelf model, integrate the subject label with the 3D archive model to obtain the 3D label archive model, and merge the 3D label archive bookshelf model and the 3D label archive model to obtain the 3D archive comprehensive model;

[0091] The dynamic display module 6 is used to update the 3D archive comprehensive model according to the adding and deleting operations of the electronic archives in the archive management system, obtain the 3D dynamic archive comprehensive model, and display it.

[0092] Embodiment 4, as Figure 6As shown, an electronic device proposed by the present invention is applicable to an electronic archive management method for project management in the tobacco industry, comprising: at least one processor 8 and a memory 7;

[0093] The memory 7 is used to store computer execution instructions;

[0094] At least one processor 8 is used to execute computer-executable instructions stored in the memory 7 to implement an electronic file management method for tobacco industry project management.

[0095] In an optional embodiment, the memory 7 can be independent or integrated with the processor 8. When the memory 7 is independently provided, the device further includes a bus 9 for connecting the memory 7 and the processor 8.

[0096] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. An electronic archive management method for tobacco industry project management, characterized in that: include: Parsing the contents of multiple electronic archives entered into the archive management system to obtain electronic archive texts, and performing preprocessing operations on the electronic archive texts to obtain standard electronic archive texts; Extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text, thereby obtaining the subject corresponding to the electronic archive, and integrating the standard electronic archive text based on the text subject to obtain an integrated electronic archive text; The integrated electronic archive text is classified by a trained text classification model to obtain the category corresponding to the integrated electronic archive text, thereby obtaining the category corresponding to the electronic archive; the text classification model includes a text representation layer, a feature extraction layer and a fully connected layer, the text representation layer adopts an ERNIE model, the feature extraction layer adopts a BiGRU model to extract global feature information, and introduces an attention mechanism into the BiGRU model, adopts a CNNs model to extract local feature information, and fuses the global feature information and the local feature information to obtain fused feature information, the fully connected layer performs dimensionality reduction processing on the fused feature information, and predicts the category of the integrated electronic archive text through a classifier; Constructing a 3D archive bookshelf model based on three-dimensional modeling technology, and constructing a 3D archive model corresponding to the electronic archive, constructing a category label of the 3D archive bookshelf model based on the category corresponding to the electronic archive, and constructing a subject label of the 3D archive model based on the subject corresponding to the electronic archive; Integrate the category label with the 3D archive bookshelf model to obtain a 3D label archive bookshelf model, integrate the subject label with the 3D archive model to obtain a 3D label archive model, and merge the 3D label archive model with the 3D label archive bookshelf model to obtain a 3D archive comprehensive model; Based on the adding and deleting operations of the plurality of electronic archives in the archive management system, the 3D archive comprehensive model is updated to obtain a 3D dynamic archive comprehensive model, and the 3D dynamic archive comprehensive model is displayed.

2. The electronic archive management method for tobacco industry project management according to claim 1, characterized in that: The extracting the subject of the standard electronic archive text to obtain the text subject corresponding to the standard electronic archive text includes: The number of topics to be generated is determined according to the size and perplexity of the standard electronic archive text. The perplexity index calculation formula is as follows: Among them, Perplexity (D) represents the perplexity of the standard electronic archive text, p (w) represents the probability of each word appearing in the standard electronic archive text, N d represents the total number of words in the dth paragraph of the standard electronic archive text, d represents the dth paragraph of the standard electronic archive text, M represents the total number of paragraphs in the standard electronic archive text, and p(w)=p(z|d)*p(w|z), p(z|d) represents the probability of the dth paragraph of the standard electronic archive text being displayed under the text theme, and p(w|z) represents the probability of a word being displayed under the text theme; Extracting the topic-word multinomial distribution parameters of the words of the text topic in the standard electronic archive text respectively according to Dirichlet distribution; Extracting the text topic distribution in the standard electronic archive text according to Dirichlet distribution based on topic-word multinomial distribution parameters to obtain the final text topic distribution and text topic word distribution; By judging the topic similarity of the word distribution in the text topic of each of the standard electronic archive texts, the topics with similar topics are merged to obtain the text topic corresponding to the standard electronic archive text. The topic similarity calculation formula is as follows: Among them, sim(T m ,T n ) represents the first text topic T m and the second text topic T n The similarity between mi and p ni Respectively represent the first text topic T m and the second text topic T n The probability distribution value of the i-th dimension topic word, v represents the first text topic T m and the second text topic T n The dimension of the words to be compared in the vector space.

3. The electronic archive management method for tobacco industry project management according to claim 1, characterized in that: The standard electronic archive text is integrated based on the text theme to obtain an integrated electronic archive text, including Creating an abstract tag table, wherein the abstract tag table includes a plurality of abstract tags, and establishing a mapping relationship table between the abstract tag table and the text topics; According to the mapping relationship table, the standard electronic archive text is mapped and matched with the abstract label table to obtain the abstract label corresponding to the standard electronic archive text; The abstract tag is matched with a preset tag content library to obtain the tag content corresponding to the abstract tag, and the tag content is added to the corresponding standard electronic archive text to obtain the integrated electronic archive text.

4. The electronic archive management method for tobacco industry project management according to claim 1, characterized in that: The BiGRU model includes a forward GRU and a reverse GRU. The forward hidden layer state calculation formula of the forward GRU at time t is: in, represents the forward hidden layer state and the reverse hidden layer state at time t-1, x t represents the input at time t, f GRU (·) represents the hidden layer calculation function of the unidirectional GRU, and the reverse hidden layer state calculation formula of the reverse GRU at time t is: in, represents the reverse hidden layer state at time t-1, and the attention mechanism calculation formula is: Among them, H t represents the feature information of the forward GRU and reverse GRU of the fusion attention mechanism at time t, f att (·) represents the calculation function of the attention mechanism, and the classifier calculation formula is y=softmax(W con F+b con ), where y represents the probability that the integrated electronic archive text belongs to all categories, softmax(·) represents the activation function, and W con represents the weight matrix, F represents the fusion feature information, b con represents the bias matrix.

5. The electronic archive management method for tobacco industry project management according to claim 1, characterized in that: The step of merging the 3D label archive model and the 3D label archive bookshelf model to obtain a 3D archive comprehensive model includes: storing all the 3D tag archive models in a list to be merged, and sorting them according to the subject tags of the 3D tag archive models; Determine whether the 3D label archive model already exists in the 3D label archive bookshelf model space; If it does not exist in the 3D label file bookshelf model space, create a strip space whose height is the same as the height of the 3D label file model, and put the 3D label file model into the strip space; If it exists in the 3D label archive bookshelf model space, create a horizontal space, and sequentially take out the 3D label archive models from the to-be-merged list and put them into the horizontal space; The strip space is placed into the 3D label archive bookshelf model space, thereby obtaining the 3D archive comprehensive model.

6. An electronic archive management system for tobacco industry project management, which is applicable to the electronic archive management method for tobacco industry project management as described in any one of claims 1 to 5, characterized in that: include: An archive parsing module (1), the archive parsing module (1) is used to import a plurality of electronic archives into an archive management system, parse the contents thereof to obtain electronic archive texts, and perform preprocessing operations to obtain standard electronic archive texts; A subject extraction module (2), the subject extraction module (2) is used to extract the subject of the standard electronic archive text to obtain the text subject, and integrate the standard electronic archive text based on the text subject to obtain the integrated electronic archive text; A text classification module (3), wherein the text classification module (3) is used to classify the text of the integrated electronic archive to obtain a category corresponding to the electronic archive; A three-dimensional modeling module (4), the three-dimensional modeling module (4) is used to construct a 3D archive bookshelf model and a 3D archive model corresponding to the electronic archive, and to construct category labels and subject labels; A label integration module (5), the label integration module (5) is used to integrate the category label with the 3D archive bookshelf model to obtain the 3D label archive bookshelf model, integrate the subject label with the 3D archive model to obtain the 3D label archive model, and merge the 3D label archive bookshelf model and the 3D label archive model to obtain the 3D archive comprehensive model; A dynamic display module (6) is used to update the 3D archive comprehensive model according to the addition and deletion operations of the electronic archive in the archive management system, obtain the 3D dynamic archive comprehensive model, and display it.

7. An electronic device, which is applicable to the electronic file management method for tobacco industry project management according to any one of claims 1 to 5, characterized in that: include: at least one processor (8) and memory (7); The memory (7) is used to store computer-executable instructions; The at least one processor (8) is used to execute the computer-executable instructions stored in the memory (7) to implement an electronic archive management method for tobacco industry project management as described in any one of claims 1 to 5.

Images