Electronic file intelligent sorting and filing method and system based on AI large model

Abstract

The invention is applicable to the technical field of file archiving, and provides an intelligent electronic file sorting and archiving method and system based on an AI large model, and the method comprises the following steps: receiving electronic text information, and carrying out title recognition and paragraph recognition on the electronic text information to obtain title information and paragraph texts; all paragraph texts are input into an AI large model for extraction description, and an extraction statement of each paragraph text is obtained; performing semantic matching on the title information and the extracted statement with classification tags, and determining classification tags corresponding to the title information and each paragraph text; the weight value of each classification label is calculated according to the title information weight and the paragraph text weight, classification label arrangement information is obtained, and N labels are arranged in the classification label arrangement information in sequence; and inputting the classification label arrangement information into the classification path identification model to obtain an archiving category and a storage path. According to the method, all contents of the paragraph text do not need to be processed, so that the consumption of computing resources is greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of file archiving, and in particular to a method and system for intelligently organizing and archiving electronic files based on an AI big model. Background Art

[0002] With the rapid development of information technology, electronic files have become the main way of information storage and transmission in modern society. Their number is growing explosively. Traditional archive classification and management methods mainly rely on manual annotation and experience-based rule formulation, which can be coped with when the number of files is small and the type is relatively single, but it seems to be unable to cope with massive and diverse electronic files. The traditional electronic file archiving classification and storage period determination process usually involves manual reading of file content, annotation according to a preset classification system, and determination of the storage period based on factors such as file type and importance. This method has the problem of low efficiency. Manual processing of files is time-consuming and labor-intensive. Especially when the number of files is large, the processing speed becomes a bottleneck, which is difficult to meet the needs of efficient management. At present, some intelligent models are used to automatically classify electronic files. However, the content of electronic files is complex and changeable, and contains rich semantic information. Existing methods are often based on keyword matching or simple rule judgment, which is difficult to accurately capture the deep semantics of the file, resulting in inaccurate classification. In addition, when the content of the document involves multiple topics or classification tags, the existing methods are difficult to accurately determine its belonging category, and often require manual intervention for the final decision. In addition, the existing methods will perform word segmentation and text feature extraction on the entire text of the file, which consumes a lot of computing resources and has a slow processing speed. Therefore, it is necessary to provide an intelligent electronic file organization and archiving method and system based on an AI big model to solve the above problems. Summary of the invention

[0003] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a method and system for intelligent organization and archiving of electronic files based on an AI big model, so as to solve the problems existing in the above-mentioned background technology.

[0004] The present invention is implemented as follows: a method for intelligently organizing and archiving electronic files based on an AI big model, the method comprising the following steps: Receiving electronic text information, performing preprocessing, performing title recognition and paragraph recognition on the electronic text information, and obtaining a title information and a plurality of paragraph texts; Input all paragraph texts into the AI ​​big model for refined description, and obtain refined sentences for each paragraph text; Semantically match the title information and the extracted sentences with the classification labels to determine the classification labels corresponding to the title information and each paragraph text; The weight value of each category label is calculated according to the title information weight and the paragraph text weight to obtain the category label arrangement information, in which N labels are arranged in order; Inputting the classification label arrangement information into the classification path recognition model to obtain the filing category and storage path, wherein the classification path recognition model is trained based on the classification labels arranged in sequence; The storage period of the electronic text information is determined according to the weight value of each classification label.

[0005] As a further solution of the present invention: the step of inputting all paragraph texts into the AI ​​big model for refining description to obtain the refined sentences of each paragraph text specifically includes: Determine the number of tokens of each paragraph text, and when the number of tokens is greater than a threshold M, segment the paragraph text. After segmentation, determine the number of tokens of each paragraph text again; Determine the number of tokens in the output sentence based on the number of tokens in the paragraph text, determine the output requirements, input the paragraph text + output requirements into the AI ​​big model, and obtain the refined sentence.

[0006] As a further solution of the present invention: the step of semantically matching the title information and the refined sentence with the classification label to determine the classification label corresponding to the title information and each paragraph text specifically includes: Retrieve the label introduction information corresponding to each classification label, and perform word segmentation on the title information, refined sentences, and label introduction information; Use bag-of-words, TF-IDF, or word embedding to extract text features from the text content after word segmentation. Semantic similarity is calculated based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

[0007] As a further solution of the present invention: the step of calculating the weight value of each classification label according to the title information weight and the paragraph text weight to obtain the classification label arrangement information specifically includes: Determine the weight value of the classification label corresponding to the title information according to the title information weight; The total weight of each paragraph is allocated according to the number of characters in each paragraph text, and the weight value of the classification label corresponding to each paragraph text is determined; The same classification labels are merged, and the classification label arrangement information is obtained based on the classification labels with the top N weight values.

[0008] As a further solution of the present invention: the step of determining the storage period of the electronic text information according to the weight value of each classification label specifically includes: Retrieve the corresponding storage time of each classification label, for a total of n classification labels; Determine the storage period H of the electronic text information according to the weight values ​​of all classification labels, ,in, Indicates the storage time of the i-th classification label, Represents the weight value of the i-th classification label.

[0009] As a further solution of the present invention: the training steps of the classification path recognition model are as follows: Collect classification label groups, each of which contains N classification labels arranged in order, and each classification label group is marked with an archiving category and a storage path; Perform one-hot encoding on the classification label group and divide it into training set, validation set and test set; Determine the input layer as a fully connected layer, add one or more convolutional layers to extract features in the classification label group, add a pooling layer after the convolutional layer, and determine the output layer, using the softmax activation function and the number of output nodes equal to the number of archive categories and storage path categories; Perform model training, determine the loss function as the cross entropy loss function, and select Adam or RMSprop as the optimizer; Use accuracy, recall, and F1 score metrics to evaluate the performance of the model on the test set and adjust the model structure or hyperparameters.

[0010] Another object of the present invention is to provide an electronic file intelligent sorting and archiving system based on an AI big model, the system comprising: The title and paragraph recognition module is used to receive electronic text information, perform preprocessing, perform title recognition and paragraph recognition on the electronic text information, and obtain a title information and a plurality of paragraph texts; The refined sentence determination module is used to input all paragraph texts into the AI ​​big model for refined description and obtain the refined sentences of each paragraph text; A classification label determination module is used to semantically match the title information and the extracted sentence with the classification label, and determine the classification label corresponding to the title information and each paragraph text; A classification label arrangement module is used to calculate the weight value of each classification label according to the title information weight and the paragraph text weight, and obtain classification label arrangement information, in which N labels are arranged in order; An archiving path determination module is used to input classification label arrangement information into a classification path recognition model to obtain an archiving category and a storage path, wherein the classification path recognition model is trained based on classification labels arranged in sequence; The storage period calculation module is used to determine the storage period of the electronic text information according to the weight value of each classification label.

[0011] As a further solution of the present invention: the sentence extraction determination module includes: A paragraph token number determination unit, used to determine the token number of each paragraph text, and when the token number is greater than a threshold value M, segment the paragraph text, and after segmentation, determine the token number of each paragraph text again; The output requirement determination unit is used to determine the number of tokens in the output sentence according to the number of tokens in the paragraph text, determine the output requirement, input the paragraph text + output requirement into the AI ​​big model, and obtain the refined sentence.

[0012] As a further solution of the present invention: the classification label determination module includes: A text segmentation processing unit is used to retrieve the label introduction information corresponding to each classification label, and perform segmentation processing on the title information, refined sentences and label introduction information; A text feature extraction unit, used to extract text features of the text content after word segmentation using a bag-of-words model, TF-IDF or word embedding; The classification label determination unit is used to calculate semantic similarity based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

[0013] As a further solution of the present invention: the classification label arrangement module includes: A first weight value calculation unit, used to determine the weight value of the classification label corresponding to the title information according to the title information weight; A second weight value calculation unit is used to allocate the total weight of each paragraph according to the number of characters in each paragraph text, and determine the weight value of the classification label corresponding to each paragraph text; The arrangement information determining unit is used to merge the same classification labels and obtain the classification label arrangement information based on the classification labels with the top N weight values.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention obtains the refined sentences of each paragraph text, and subsequently only needs to analyze and process the refined sentences, without processing the entire content of the paragraph text, which greatly reduces the consumption of computing resources. The present invention calculates the weight value of each classification label according to the title information weight and the paragraph text weight, obtains the classification label arrangement information, and inputs the classification label arrangement information into the classification path recognition model, which will automatically output the filing category and storage path. In this way, when a file involves multiple classification labels, its belonging category can also be accurately judged to improve the rationality of the classification. The storage life of the electronic text information is also determined according to the weight value of each classification label. Through the automated processing capability of the AI ​​large model, rapid classification and archiving of massive electronic files can be achieved, significantly improving work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The present invention is a flowchart of an intelligent electronic file organization and archiving method based on an AI big model.

[0016] Figure 2 The present invention is a flowchart for determining and refining sentences in an intelligent electronic file organization and archiving method based on an AI big model.

[0017] Figure 3 The present invention is a flowchart for determining classification labels in an intelligent organization and archiving method for electronic files based on an AI big model.

[0018] Figure 4 The present invention is a flowchart for obtaining classification label arrangement information in an intelligent electronic file organization and archiving method based on an AI big model.

[0019] Figure 5 The present invention is a flowchart for determining the storage period in an intelligent organization and archiving method for electronic files based on an AI big model.

[0020] Figure 6 The present invention is a flowchart for training a classification path recognition model in an intelligent electronic file organization and archiving method based on an AI big model.

[0021] Figure 7 This is a structural diagram of an intelligent electronic file organization and archiving system based on an AI big model. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0023] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0024] like Figure 1 As shown, an embodiment of the present invention provides an electronic file intelligent sorting and archiving method based on an AI big model, and the method includes the following steps: S100, receiving electronic text information, performing preprocessing, performing title recognition and paragraph recognition on the electronic text information, and obtaining a title information and a plurality of paragraph texts; S200, input all paragraph texts into the AI ​​big model for extraction and description, and obtain the extracted sentences of each paragraph text; S300, semantically matching the title information and the extracted sentence with the classification label to determine the classification label corresponding to the title information and each paragraph text; S400, calculating the weight value of each classification label according to the title information weight and the paragraph text weight, and obtaining classification label arrangement information, wherein N labels are arranged in sequence in the classification label arrangement information; S500, inputting classification label arrangement information into a classification path recognition model to obtain an archiving category and a storage path, wherein the classification path recognition model is trained based on classification labels arranged in sequence; S600: Determine the storage period of the electronic text information according to the weight value of each classification label.

[0025] In an embodiment of the present invention, after receiving the electronic text information uploaded by the user, preprocessing will be performed first, and the preprocessing includes format unification and data cleaning. Then, the electronic text information will be subject to title recognition and paragraph recognition to obtain a title information and several paragraph texts; then all the paragraph texts are input into the AI ​​big model for refined description to obtain the refined sentences of each paragraph text. The AI ​​big model can use ChatGPT, Deepseek, Tongyi Qianwen, etc., and only the refined sentences need to be analyzed and processed later, without the need to process the entire content of the paragraph text, which greatly reduces the consumption of computing resources. Then, the title information and the refined sentences are semantically matched with the classification labels to determine the classification labels corresponding to the title information and each paragraph text, and which specific classification labels need to be set in advance. Then, the weight value of each classification label is calculated according to the title information weight and the paragraph text weight, for example, the title information weight is 30%, and the sum of the weights of all paragraph texts is 70%, and the classification label arrangement information is obtained. In the classification label arrangement information, N labels are arranged in order, for example, N is five. Then, the classification label arrangement information is input into the classification path recognition model, and the archiving category and storage path will be automatically output. The classification path recognition model is trained in advance based on the classification labels arranged in sequence. The content and arrangement order of N labels will affect the final output result. In this way, when a file involves multiple classification labels, its category can be accurately determined to improve the rationality of classification. Finally, the storage period of the electronic text information will be determined according to the weight value of each classification label, which significantly improves work efficiency.

[0026] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of inputting all paragraph texts into the AI ​​big model for refining description to obtain the refined sentence of each paragraph text specifically includes: S201, determining the number of tokens of each paragraph text, and when the number of tokens is greater than a threshold value M, segmenting the paragraph text, and after segmentation, determining the number of tokens of each paragraph text again; S202, determine the number of tokens of the output sentence according to the number of tokens in the paragraph text, determine the output requirements, input the paragraph text + output requirements into the AI ​​big model, and obtain the refined sentence.

[0027] In an embodiment of the present invention, before inputting a paragraph text into the AI ​​big model, it is necessary to determine the number of tokens of each paragraph text. When the number of tokens is greater than a threshold value M, for example, M=4000, the paragraph text is segmented. When segmenting, the number of segments is determined by dividing the number of tokens by M. After segmentation, the number of tokens of each paragraph text is determined again. Then, the number of tokens of the output sentence is determined based on the number of paragraph text tokens. For example, the number of tokens of the output sentence = the number of paragraph text tokens / 4. The number of tokens of the output sentence has a minimum limit, for example, 50. Then, the output requirement is determined, and the paragraph text + output requirement is input into the AI ​​big model to obtain a refined sentence. For example, the number of tokens of the output sentence corresponding to a paragraph text is 337, and 337 / 4 is equal to 84.25. The output requirement is: refine the content of the paragraph text and output an answer content of about 84 tokens.

[0028] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of semantically matching the title information and the refined sentence with the classification label to determine the classification label corresponding to the title information and each paragraph text specifically includes: S301, retrieve the label introduction information corresponding to each classification label, and perform word segmentation processing on the title information, refined sentences and label introduction information; S302, extracting text features of the text content after word segmentation using a bag-of-words model, TF-IDF or word embedding; S303, performing semantic similarity calculation based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

[0029] In an embodiment of the present invention, many classification labels are formulated, and each classification label corresponds to label introduction information. First, the label introduction information of the classification label corresponding to the title information and each paragraph text is retrieved, and then the title information, the refined sentence and the label introduction information are segmented, and the WordPiece segmentation method is used here. Then, the bag of words model, TF-IDF or word embedding is used to extract the text features of the text content after the segmentation process. These features should be able to reflect the theme, keywords, semantics and other information of the text, and then the semantic similarity calculation is performed based on the text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

[0030] like Figure 4 As shown, as a preferred embodiment of the present invention, the step of calculating the weight value of each classification label according to the title information weight and the paragraph text weight to obtain the classification label arrangement information specifically includes: S401, determining a weight value of a classification label corresponding to the title information according to the title information weight; S402, allocating the total weight of each paragraph according to the number of characters in each paragraph text, and determining the weight value of the classification label corresponding to each paragraph text; S403, merging the same classification labels, and obtaining classification label arrangement information based on the classification labels with the top N weight values.

[0031] In the embodiment of the present invention, the weight value of the classification label corresponding to the title information is first determined according to the weight of the title information, for example, 30% here. Then, the total weight of the paragraph is allocated according to the number of characters in each paragraph text, and the total weight of the paragraph is 70%. After the total weight allocation is completed, the weight value of the classification label corresponding to each paragraph text can be determined. The same classification labels need to be merged, and finally the classification labels with weight values ​​in the top N are determined to obtain the classification label arrangement information, where N is a fixed value. For example, the classification label arrangement information is: label 26, label 41, label 12, label 87, and label 63; when there are less than N classification labels, all of them are arranged.

[0032] like Figure 5 As shown, as a preferred embodiment of the present invention, the step of determining the storage period of the electronic text information according to the weight value of each classification label specifically includes: S601, retrieve the corresponding storage time of each classification tag, a total of n classification tags; S602: Determine the storage period H of the electronic text information according to the weight values ​​of all classification tags.

[0033] In the embodiment of the present invention, each classification tag is set with a storage time in advance. In order to determine the storage period of the electronic text information, the storage period H of the electronic text information is determined according to the weight values ​​of all the classification tags of the electronic text information. ,in, Indicates the storage time of the i-th classification label, represents the weight value of the i-th classification label, .

[0034] like Figure 6 As shown, as a preferred embodiment of the present invention, the training steps of the classification path recognition model are as follows: S011, collecting classification label groups, each classification label group includes N classification labels arranged in order, and each classification label group is marked with an archiving category and a storage path; S012, perform one-hot encoding on the classification label group and divide it into training set, validation set and test set; S013, determining that the input layer is a fully connected layer, adding one or more convolutional layers to extract features in the classification label group, adding a pooling layer after the convolutional layer, and determining the output layer, using a softmax activation function and a number of output nodes equal to the number of archive categories and storage path categories; S014, perform model training, determine the loss function as the cross entropy loss function, and select Adam or RMSprop as the optimizer; S015, use accuracy, recall, and F1 score indicators to evaluate the performance of the model on the test set and adjust the model structure or hyperparameters.

[0035] In the embodiment of the present invention, the classification path recognition model needs to be trained in advance. First, a large number of classification label groups need to be collected, each of which contains N or less classification labels arranged in order, and each classification label group is marked with an archiving category and a storage path. Then each classification label is converted into a numerical form, using unique hot encoding, and the data set is divided into a training set, a validation set, and a test set. Then the CNN model is constructed: the input layer is determined to be a fully connected layer, and one or more convolutional layers are added to extract features from the classification label group. The size and number of convolutional kernels can be adjusted according to data characteristics and task requirements. Activation functions such as ReLU are used to increase nonlinearity. A pooling layer is added after the convolutional layer to reduce the dimension of the feature map, reduce the amount of calculation, and enhance the robustness of the model; the output layer is determined, and a softmax activation function and an output node number equal to the number of archiving categories and storage path categories are used. Then, we will train the model: determine the loss function as the cross entropy loss function, and select Adam or RMSprop as the optimizer; use the training set data to train the model, adjust the model parameters through back propagation, and use the validation set to evaluate the model performance at the end of each epoch, and adjust the learning rate, batch size and other hyperparameters as needed. Finally, use the accuracy, recall and F1 score indicators to evaluate the performance of the model on the test set, and adjust the model structure (such as the size and number of convolution kernels, the number of nodes in the fully connected layer, etc.) or hyperparameters (such as the learning rate, batch size, number of training rounds, etc.) according to the evaluation results.

[0036] like Figure 7 As shown, an embodiment of the present invention further provides an electronic file intelligent sorting and archiving system based on an AI big model, the system comprising: The title and paragraph recognition module 100 is used to receive electronic text information, perform preprocessing, perform title recognition and paragraph recognition on the electronic text information, and obtain a title information and a plurality of paragraph texts; The refined sentence determination module 200 is used to input all paragraph texts into the AI ​​big model for refined description to obtain the refined sentence of each paragraph text; The classification label determination module 300 is used to semantically match the title information and the refined sentence with the classification label to determine the classification label corresponding to the title information and each paragraph text; The classification label arrangement module 400 is used to calculate the weight value of each classification label according to the title information weight and the paragraph text weight to obtain classification label arrangement information, in which N labels are arranged in order; The filing path determination module 500 is used to input the classification label arrangement information into the classification path recognition model to obtain the filing category and storage path, wherein the classification path recognition model is trained based on the classification labels arranged in sequence; The storage period calculation module 600 is used to determine the storage period of the electronic text information according to the weight value of each classification label.

[0037] As a preferred embodiment of the present invention, the sentence extraction determination module 200 includes: A paragraph token number determination unit, used to determine the token number of each paragraph text, and when the token number is greater than a threshold value M, segment the paragraph text, and after segmentation, determine the token number of each paragraph text again; The output requirement determination unit is used to determine the number of tokens in the output sentence according to the number of tokens in the paragraph text, determine the output requirement, input the paragraph text + output requirement into the AI ​​big model, and obtain the refined sentence.

[0038] As a preferred embodiment of the present invention, the classification label determination module 300 includes: A text segmentation processing unit is used to retrieve the label introduction information corresponding to each classification label, and perform segmentation processing on the title information, refined sentences and label introduction information; A text feature extraction unit, used to extract text features of the text content after word segmentation using a bag-of-words model, TF-IDF or word embedding; The classification label determination unit is used to calculate semantic similarity based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

[0039] As a preferred embodiment of the present invention, the classification label arrangement module 400 includes: A first weight value calculation unit, used to determine the weight value of the classification label corresponding to the title information according to the title information weight; A second weight value calculation unit is used to allocate the total weight of each paragraph according to the number of characters in each paragraph text, and determine the weight value of the classification label corresponding to each paragraph text; The arrangement information determining unit is used to merge the same classification labels and obtain the classification label arrangement information based on the classification labels with the top N weight values.

[0040] The above only describes in detail the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0041] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0042] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0043] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A method for intelligently organizing and archiving electronic files based on an AI big model, characterized in that: The method comprises the following steps: Receiving electronic text information, performing preprocessing, performing title recognition and paragraph recognition on the electronic text information, and obtaining a title information and a plurality of paragraph texts; Input all paragraph texts into the AI ​​big model for refined description, and obtain refined sentences for each paragraph text; Semantically match the title information and the extracted sentences with the classification labels to determine the classification labels corresponding to the title information and each paragraph text; The weight value of each category label is calculated according to the title information weight and the paragraph text weight to obtain the category label arrangement information, in which N labels are arranged in order; Inputting the classification label arrangement information into the classification path recognition model to obtain the filing category and storage path, wherein the classification path recognition model is trained based on the classification labels arranged in sequence; The storage period of the electronic text information is determined according to the weight value of each classification label.

2. The method for intelligently organizing and archiving electronic files based on the AI ​​big model according to claim 1 is characterized in that: The step of inputting all paragraph texts into the AI ​​big model for refining description to obtain the refined sentences of each paragraph text specifically includes: Determine the number of tokens of each paragraph text, and when the number of tokens is greater than a threshold M, segment the paragraph text. After segmentation, determine the number of tokens of each paragraph text again; Determine the number of tokens in the output sentence based on the number of tokens in the paragraph text, determine the output requirements, input the paragraph text + output requirements into the AI ​​big model, and obtain the refined sentence.

3. The method for intelligently organizing and archiving electronic files based on the AI ​​big model according to claim 1 is characterized in that: The step of semantically matching the title information and the refined sentence with the classification label to determine the classification label corresponding to the title information and each paragraph text specifically includes: Retrieve the label introduction information corresponding to each classification label, and perform word segmentation on the title information, refined sentences, and label introduction information; Use bag-of-words, TF-IDF, or word embedding to extract text features from the text content after word segmentation. Semantic similarity is calculated based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

4. The method for intelligently organizing and archiving electronic files based on the AI ​​big model according to claim 1 is characterized in that: The step of calculating the weight value of each classification label according to the title information weight and the paragraph text weight to obtain the classification label arrangement information specifically includes: Determine the weight value of the classification label corresponding to the title information according to the title information weight; The total weight of each paragraph is allocated according to the number of characters in each paragraph text, and the weight value of the classification label corresponding to each paragraph text is determined; The same classification labels are merged, and the classification label arrangement information is obtained based on the classification labels with the top N weight values.

5. The method for intelligently organizing and archiving electronic files based on the AI ​​big model according to claim 1 is characterized in that: The step of determining the storage period of the electronic text information according to the weight value of each classification label specifically includes: Retrieve the corresponding storage time of each classification label, for a total of n classification labels; Determine the storage period H of the electronic text information according to the weight values ​​of all classification labels, ,in, Indicates the storage time of the i-th classification label, Represents the weight value of the i-th classification label.

6. The method for intelligently organizing and archiving electronic files based on the AI ​​big model according to claim 1 is characterized in that: The training steps of the classification path recognition model are as follows: Collect classification label groups, each of which contains N classification labels arranged in order, and each classification label group is marked with an archiving category and a storage path; Perform one-hot encoding on the classification label group and divide it into training set, validation set and test set; Determine the input layer as a fully connected layer, add one or more convolutional layers to extract features in the classification label group, add a pooling layer after the convolutional layer, and determine the output layer, using the softmax activation function and the number of output nodes equal to the number of archive categories and storage path categories; Perform model training, determine the loss function as the cross entropy loss function, and select Adam or RMSprop as the optimizer; Use accuracy, recall, and F1 score metrics to evaluate the performance of the model on the test set and adjust the model structure or hyperparameters.

7. An intelligent electronic file sorting and archiving system based on AI big model, characterized in that: The system comprises: The title and paragraph recognition module is used to receive electronic text information, perform preprocessing, perform title recognition and paragraph recognition on the electronic text information, and obtain a title information and a plurality of paragraph texts; The refined sentence determination module is used to input all paragraph texts into the AI ​​big model for refined description and obtain the refined sentences of each paragraph text; A classification label determination module is used to semantically match the title information and the extracted sentence with the classification label, and determine the classification label corresponding to the title information and each paragraph text; A classification label arrangement module is used to calculate the weight value of each classification label according to the title information weight and the paragraph text weight, and obtain classification label arrangement information, in which N labels are arranged in order; An archiving path determination module is used to input classification label arrangement information into a classification path recognition model to obtain an archiving category and a storage path, wherein the classification path recognition model is trained based on classification labels arranged in sequence; The storage period calculation module is used to determine the storage period of the electronic text information according to the weight value of each classification label.

8. The electronic file intelligent sorting and archiving system based on AI big model according to claim 7 is characterized in that: The sentence extraction determination module comprises: A paragraph token number determination unit, used to determine the token number of each paragraph text, and when the token number is greater than a threshold value M, segment the paragraph text, and after segmentation, determine the token number of each paragraph text again; The output requirement determination unit is used to determine the number of tokens in the output sentence according to the number of tokens in the paragraph text, determine the output requirement, input the paragraph text + output requirement into the AI ​​big model, and obtain the refined sentence.

9. The electronic file intelligent sorting and archiving system based on AI big model according to claim 7 is characterized in that: The classification label determination module includes: A text segmentation processing unit is used to retrieve the label introduction information corresponding to each classification label, and perform segmentation processing on the title information, refined sentences and label introduction information; A text feature extraction unit, used to extract text features of the text content after word segmentation using a bag-of-words model, TF-IDF or word embedding; The classification label determination unit is used to calculate semantic similarity based on text features and the BERT model to determine the classification label corresponding to the title information and each paragraph text.

10. The electronic file intelligent sorting and archiving system based on AI big model according to claim 7 is characterized in that: The classification label arrangement module includes: A first weight value calculation unit, used to determine the weight value of the classification label corresponding to the title information according to the title information weight; A second weight value calculation unit is used to allocate the total weight of each paragraph according to the number of characters in each paragraph text, and determine the weight value of the classification label corresponding to each paragraph text; The arrangement information determining unit is used to merge the same classification labels and obtain the classification label arrangement information based on the classification labels with the top N weight values.

Images