Intelligent archive management method and system based on big data

Abstract

The invention relates to the technical field of archive management, in particular to an intelligent archive management method and system based on big data. A smart archive management method based on big data comprises the following steps: S1, obtaining related data of a paper archive and an electronic backup archive, obtaining a predicted recording time result of the paper archive by using a recording time detection model based on the related data, and storing the predicted recording time result in the electronic backup archive; and S2, based on the basic information of each paper archive in the electronic backup archive and the predicted recording time data, judging whether a recording time requirement is met, and performing archive management. By utilizing the word embedding and adjusting formula, the keyword can be subjected to vectorization mapping in combination with the historical query data of the user, the query habit and the information requirement of the user are analyzed, and the query range is dynamically adjusted according to the access frequency, so that the query result better meets the user requirement, and the use experience of the user is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of archive management, and in particular to a smart archive management method and system based on big data. Background Art

[0002] Traditional archive management mainly relies on manual records and simple electronic data storage methods, lacking intelligent management methods, resulting in low efficiency, high labor costs and insufficient data accuracy in the process of archive collection, query and update. Especially in paper archive management, there is uncertainty in the estimation of collection time and archiving work, which cannot meet the needs of large-scale data management. In addition, the existing archive management system is difficult to effectively meet the diverse query needs of users, and it is easy to omit content or redundant information during information retrieval, affecting user experience and management efficiency; when using their own commonly used keys, it is easy to leak their own keys, so it will become very dangerous to query archives after the key is leaked.

[0003] With the development of technologies such as big data, artificial intelligence and natural language processing, intelligent and smart file management has gradually become a trend. By applying advanced technical means such as big data analysis technology, machine learning algorithms and user behavior analysis, intelligent solutions can be provided for file management. Summary of the invention

[0004] In order to overcome the shortcomings of personnel leaking common keys and incomplete queries when querying archives, the present invention provides a smart archive management method and system based on big data.

[0005] The technical implementation scheme of the present invention is: a smart archive management method based on big data, comprising the following steps: S1: Obtain relevant data of paper archives and electronic backup files, and use the collection time detection model based on the relevant data to obtain the predicted collection time result of the paper archives, and save the predicted collection time result to the electronic backup file; S2: Based on the basic information of each paper archive in the electronic archive and the predicted collection time data, determine whether the collection time requirement is met and perform archive management; S3: Acquire access data and query keyword data of the target user, and determine the query content scope of the target user according to the access data and query keyword data of the target user; S4: using the bag-of-words model to convert the summary content of each file in the database into a summary vector, generating a query value for each file according to the summary vector, and filtering out a preset number of files as query results according to the key type of the target user; S5: Based on the query content, the target user is displayed and an option for the user to log in again for the query is displayed.

[0006] Preferably, the method of obtaining relevant data of paper archives and electronic backup files, and using a collection time detection model based on the relevant data to obtain a predicted collection time result of the paper archives, and saving the predicted collection time result in the electronic backup file includes: collecting the collection time, paper material, pen and ink quality, storage environment data and software recognition success rate data of paper archives in different batches, and performing data cleaning to remove outliers and fill in missing values ​​to obtain pre-processed data, and inputting the pre-processed data into the collection time detection model to obtain the predicted collection time of each batch of paper archives; wherein the collection time detection model is a Transformer model trained using historical data and optimized using a cross-validation method; the electronic backup file contains basic information of paper archives that are not included in the electronic archive and the predicted collection time data.

[0007] Preferably, based on the basic information of each paper archive in the electronic backup file and the predicted collection time data, it is judged whether the collection time requirement is met, and archive management is performed, including: judging whether each batch of paper archives will meet the collection time requirement, if the predicted time data of the paper archive is within the first preset time of the current time, the paper archive is added to the archive update sequence; and according to the protection of each paper archive in the archive update sequence, the electronic backup file is updated or entered into the electronic archive.

[0008] Preferably, the acquiring of the target user's access data and query keyword data, and determining the target user's query content range based on the target user's access data and query keyword data, includes: acquiring the target user's query keyword data, and using word embedding technology to map the keyword data to obtain keyword vectors, and determining a query range adjustment value based on each keyword vector and the target user's access data; and obtaining the target user's query content range based on the query range adjustment value.

[0009] Preferably, determining the query range adjustment value according to each keyword vector and the access data of the target user includes: obtaining the historical query times of each keyword in the access data of the target user and each keyword vector, and using a query range adjustment formula to obtain the query range adjustment value, wherein the adjustment formula is: ;in is the adjustment value of the query range, is the total number of search keywords, For the query keyword keyword vectors, For keywords The number of historical queries, It is the total number of occurrences of all keywords in historical query records.

[0010] Preferably, the query content range of the target user is obtained according to the query range adjustment value, including: obtaining the maximum number of accessible archives of the target user according to the key type input by the target user, and obtaining the query content range of the target user according to the maximum number of accessible archives using an upper limit calculation formula and a lower limit calculation formula, wherein the upper limit calculation formula is: ; In the formula, is the adjustment value of the query range, is the total number of query keywords, T is the maximum number of accessible archives for the target user; The lower limit calculation formula is: .

[0011] Preferably, the method of converting the summary content of each file in the database into a summary vector using a bag-of-words model, generating a query value for each file according to the summary vector, and selecting a preset number of files as query results according to the key type of the target user includes: obtaining the key type of the target user when logging in; If it is the first key, the summary content of each file in the database and the summary content in the electronic backup file are converted into a summary vector using the bag-of-words model, and a query value is generated for each file according to the summary vector. According to the query content range and query value of the target user, the files whose query values ​​are within the query content range and greater than or equal to the preset threshold are selected as the query results; If it is the second key, the bag-of-words model is used to convert the summary content of each file in the database into a summary vector, and a query value is generated for each file based on the summary vector. The query content range of the target user is adjusted using the security adjustment formula to obtain an adjusted query content range. Based on the adjusted query content range and the query value, the files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results.

[0012] Preferably, if the first key is used, the summary content of each file in the database and the summary content in the electronic backup file are converted into a summary vector using a bag-of-words model, and a query value is generated for each file according to the summary vector. According to the query content range and query value of the target user, the files whose query values ​​are within the query content range and are greater than or equal to a preset threshold are selected as query results, including: the first key and the second key are the login keys of the target user, the first key is the standard login key; the second key is the login key set by the target user, and the query value of each file is obtained through the query value calculation formula, wherein the query value calculation formula is: ; In the formula is the query value, is the summary vector corresponding to the summary content of the archive, The total number of words in the abstract content in the archive.

[0013] Preferably, if the second key is used, the summary content of each file in the database is converted into a summary vector using a bag-of-words model, and a query value is generated for each file according to the summary vector, and the query content range of the target user is adjusted using a security adjustment formula to obtain an adjusted query content range, and according to the adjusted query content range and the query value, the files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results, including: adjusting the adjustment value of the query range using a security adjustment formula to obtain a second adjustment value, and obtaining the adjusted query content range according to the second adjustment value, wherein the security adjustment formula is: ; In the formula, is the second adjustment value and Greater than or equal to 0, is the adjustment value of the query range, is the adjustment coefficient, The number of times the second key was used to log in after the first key was used to log in most recently. The maximum number of times tolerable to log in using the second key.

[0014] Preferably, a smart archive management system based on big data includes: The data collection and preprocessing module is responsible for collecting basic information data of paper archives and electronic backup files, including the collection time, paper material, pen and ink quality, and storage environment of the archives, and performs data denoising, outlier processing, and missing value filling cleaning operations; The collection time prediction module trains the preprocessed archive data based on the collection time detection model, predicts the collection time of each batch of paper archives, and saves the prediction results to the electronic backup file; The archive management module manages archives according to the predicted collection time and archive protection, and determines whether the collection time requirement is met. If the conditions are met, the archive is added to the archive update sequence, and then the electronic backup file is updated or entered into the electronic archive; The user access data analysis module is responsible for obtaining and analyzing the access data and query keyword data of the target user, converting the keyword into a vector using word embedding technology, and determining the query range adjustment value based on the user's historical access data and keyword vector, so as to provide the user with a reasonable query content range in the subsequent steps; The query authority and display module determines the query content range of the user according to the first key or the second key, selects the files whose query values ​​are within the query content range and are greater than or equal to a preset threshold as the query results, and displays them.

[0015] The beneficial effects of the present invention are as follows: the present invention collects relevant data of paper archives and electronic backup archives and applies the collection time detection model to intelligently predict the collection time of each batch of paper archives, effectively avoids the deviation of manual judgment, makes archive management more efficient and accurate, and reduces the need for manual intervention; Obtain the target user's access data and query keyword data, use the bag-of-words model to generate a summary vector, and combine it with the query range adjustment value formula to enable the system to dynamically determine the query content range based on user behavior data. This not only improves query accuracy, but also avoids information redundancy and improves user experience; Through the distinction of key types and the query range screening mechanism corresponding to different keys, the access scope of the query content is dynamically controlled according to user permissions and needs. This method ensures the query experience of users under different permissions, while improving the security of archive data and ensuring access control of important information; This system uses word embedding and adjustment formulas to combine users' historical query data, perform vectorized mapping of keywords, analyze users' query habits and information needs, and dynamically adjust the query range according to the access frequency, so that the query results are more in line with user needs and enhance the user experience; A tolerance upper limit is set for the number of times the second login key can be used, and the scope of query content is adjusted according to user access conditions, effectively improving the security of data access. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a smart archive management method based on big data of the present invention; Figure 2 This is a relationship diagram of a smart archive management system based on big data in the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Embodiment 1: A smart archive management method based on big data, such as Figure 1-Figure 2 As shown, the following steps are included: S1: Obtain relevant data of paper archives and electronic backup files, and use the collection time detection model based on the relevant data to obtain the predicted collection time result of the paper archives, and save the predicted collection time result to the electronic backup file; The collection time, paper material, pen and ink quality, storage environment data and software recognition success rate data of paper archives in different batches are collected, and the pre-processed data are obtained after data cleaning to remove outliers and fill in missing values. The pre-processed data is input into the collection time detection model to obtain the predicted collection time of each batch of paper archives; wherein, the collection time detection model is a Transformer model trained using historical data and tuned using the cross-validation method; the electronic backup file contains the basic information of the paper archives that are not included in the electronic archives and the predicted collection time data.

[0019] It should be noted that the data of different batches of paper archives are collected, including the collection time: the actual collection time of different batches of paper archives for model prediction; paper material: the material characteristics of the paper used for different archives, pH, thickness and fiber structure; pen and ink quality: record the type, quality and color depth of the pen and ink used in the archives, so that the model can analyze the aging speed of the archival materials; storage environment data: temperature, humidity and light conditions of the environment where the archives are located. Software recognition success rate data: recognition success rate data of the recognition software (such as OCR software) used in electronicization; the collected data is cleaned to remove outliers and fill in missing values. Remove outliers: remove data points with extreme deviations, such as abnormally low recognition success rates; fill in missing values: supplement the missing collection time and material information by interpolation to ensure data integrity. The preprocessed data is input into the collection time detection model. The model uses the Transformer structure and is trained and cross-validated with historical data. After tuning, the model can generate reliable collection time prediction results. The predicted collection time results of each batch of paper archives are saved in the electronic backup file. Electronic archives record the basic information of uncollected paper archives and predicted collection time data to facilitate subsequent archive management operations.

[0020] S2: Based on the basic information of each paper archive in the electronic archive and the predicted collection time data, determine whether the collection time requirement is met and perform archive management; Determine whether each batch of paper archives will meet the collection time requirements. If the predicted time data of the paper archives is within the first preset time of the current time, the paper archives will be added to the archive update sequence; and according to the protection of each paper archive in the archive update sequence, update the electronic backup file or enter it into the electronic archive.

[0021] It should be noted that based on the basic information of the archives in the electronic backup and the predicted collection time, the system determines whether the archive is about to reach the collection time requirement. The judgment criteria are: if the predicted time data of the archive is within the first preset time of the current time (for example, 3 months), it is considered that the archive is about to reach the collection time; the system will add the paper archives that meet the collection time requirements to the archive update sequence for priority processing; the system evaluates the protection of the archives (such as damage or importance) based on the archive information in the archive update sequence, and chooses whether to update the electronic backup or directly enter the electronic archive. For example, some archives are very important and should be kept in paper archives. At this time, the paper archives should be updated instead of being entered into the electronic archives. If some archives are highly damaged and have reduced confidentiality, they should be entered into the electronic archives to avoid information loss due to paper damage.

[0022] S3: Acquire access data and query keyword data of the target user, and determine the query content scope of the target user according to the access data and query keyword data of the target user; Obtain the query keyword data of the target user, and use word embedding technology to map the keyword data to obtain keyword vectors, determine the query range adjustment value based on each keyword vector and the target user's access data; and obtain the query content range of the target user based on the query range adjustment value.

[0023] It should be noted that the system first obtains the user's query keywords and their access records. The user's access records include the number of search times for the query keywords, the frequency of keyword use, the use of the first key or the second key, and various key types. The system vectorizes the obtained query keyword data through word embedding technology. Word embedding technology converts text keywords into numerical vectors, thereby calculating the semantic distance between keywords.

[0024] Obtain the historical query times of each keyword and each keyword vector in the target user's access data, and use the query range adjustment formula to obtain the adjustment value of the query range, where the adjustment formula is: ;

[0025] in is the adjustment value of the query range, is the total number of search keywords, For the query keyword keyword vectors, For keywords The number of historical queries, It is the total number of occurrences of all keywords in historical query records.

[0026] According to the key type input by the target user, the maximum number of accessible files of the target user is obtained, and according to the maximum number of accessible files, the upper limit calculation formula and the lower limit calculation formula are used to obtain the query content range of the target user, wherein the upper limit calculation formula is: ; In the formula, is the adjustment value of the query range, is the total number of query keywords, T is the maximum number of accessible archives for the target user; The lower limit calculation formula is: It should be noted that the use of , the higher this value is, the more frequently users search for this type of keyword. When users are interested in this type of keyword, the query content range is relaxed, and more content related to this keyword content is retrieved. When there are many keywords, it will no longer be a determining factor; the system determines the maximum number of accessible archives for the user based on the key type entered by the user (such as ordinary user or advanced user). The key type can reflect the user's authority range. Different key types correspond to different numbers of accessible archives T. The system selects qualified archives based on the upper and lower limits of the query content, and displays them to the user based on the query content priority. This operation ensures that users can quickly find archive content that meets their needs, while ensuring the security of the system and user authority.

[0027] S4: using the bag-of-words model to convert the summary content of each file in the database into a summary vector, generating a query value for each file according to the summary vector, and filtering out a preset number of files as query results according to the key type of the target user; Get the key type of the target user when logging in; If it is the first key, the summary content of each file in the database and the summary content in the electronic backup file are converted into a summary vector using the bag-of-words model, and a query value is generated for each file according to the summary vector. According to the query content range and query value of the target user, the files whose query values ​​are within the query content range and greater than or equal to the preset threshold are selected as the query results; If it is the second key, the bag-of-words model is used to convert the summary content of each file in the database into a summary vector, and a query value is generated for each file based on the summary vector. The query content range of the target user is adjusted using the security adjustment formula to obtain an adjusted query content range. Based on the adjusted query content range and the query value, the files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results.

[0028] It should be noted that the first key: as a standard login key, has high authority and a wide search range, ensuring that users have comprehensive access to archives that meet the query criteria; the second key: is set by the user, has low authority, and because of its low security, it is easy to be stolen or lost. Therefore, the system will limit the search scope to ensure the security of the archive data.

[0029] When using the first key, the system converts the summary content of each file in the database and the summary content in the electronic backup file into a corresponding summary vector. The system generates a query value based on the summary vector of each file, which is used to indicate the degree of match between the file content and the user's query. After generating the query value, the system will filter out files whose query values ​​meet the requirements based on the query content range of the first key user. Only those files whose query values ​​fall within the query content range and are greater than the preset threshold can be selected as query results and displayed to the user.

[0030] When using the second key, the system only converts the summary content of each file in the database into a corresponding summary vector, generates a query value based on the summary vector of each file, and the query value is used to indicate the degree of match between the file content and the user query. After adjusting the query content range of the target user using the security adjustment formula, an adjusted query content range is obtained. Based on the adjusted query content range and the query value, files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results.

[0031] The first key and the second key are the login keys of the target user, the first key is the standard login key; the second key is the login key set by the target user, and the query value of each file is obtained through the query value calculation formula, wherein the query value calculation formula is: ; In the formula is the query value, is the summary vector corresponding to the summary content of the archive, The total number of words in the abstract content in the archive.

[0032] It should be noted that the first key and the second key are both login keys of the target user, the first key is a standard login key; the second key is a login key set by the target user, so the first key and the second key have different key types. According to the key type of the first key or the second key, the number T of accessible files for different user identities is obtained.

[0033] The adjustment value of the query range is adjusted using a security adjustment formula to obtain a second adjustment value, and the adjusted query content range is obtained according to the second adjustment value, wherein the security adjustment formula is: ; In the formula, is the second adjustment value and Greater than or equal to 0, is the adjustment value of the query range, is the adjustment coefficient, The number of times the second key was used to log in after the first key was used to log in most recently. The maximum number of times tolerable to log in using the second key.

[0034] It should be noted that when a user logs in with the second key, due to the low authority and limited search range, the system will limit the scope of query content based on the number of times the user logs in with the second key to reduce the risk of data leakage. Security adjustment formula: The system will limit the query range of the second key user through a formula, making the query range of the second key smaller than the first key. By controlling the access range with different key permissions, the first key user can obtain more comprehensive query results, while the second key user can only see limited data after security adjustment; for the second key that is easily lost or stolen, the system automatically narrows the query range to effectively protect the security of the archive data. It is the number of times the second key was used to log in after the first key was used most recently. Therefore, after logging in again with the first key, the query range of the archive will become larger.

[0035] S5: Based on the query content, the target user is displayed and an option for the user to log in again for the query is displayed.

[0036] It should be noted that if the target user feels that the content query is not comprehensive, a re-login query option is provided for the user to log in using the first key to increase the query scope of the archive.

[0037] Embodiment 2: Based on Embodiment 1, a smart archive management system based on big data includes: The data collection and preprocessing module is responsible for collecting basic information data of paper archives and electronic backup files, including the collection time, paper material, pen and ink quality, and storage environment of the archives, and performs data denoising, outlier processing, and missing value filling cleaning operations; The collection time prediction module trains the preprocessed archive data based on the collection time detection model, predicts the collection time of each batch of paper archives, and saves the prediction results to the electronic backup file; The archive management module manages archives according to the predicted collection time and archive protection, and determines whether the collection time requirement is met. If the conditions are met, the archive is added to the archive update sequence, and then the electronic backup file is updated or entered into the electronic archive; The user access data analysis module is responsible for obtaining and analyzing the access data and query keyword data of the target user, converting the keyword into a vector using word embedding technology, and determining the query range adjustment value based on the user's historical access data and keyword vector, so as to provide the user with a reasonable query content range in the subsequent steps; The query authority and display module determines the query content range of the user according to the first key or the second key, selects the files whose query values ​​are within the query content range and are greater than or equal to a preset threshold as the query results, and displays them.

[0038] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes may be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart archive management method based on big data, characterized in that: The following steps are involved: S1: Obtain relevant data of paper archives and electronic backup files, and use the collection time detection model based on the relevant data to obtain the predicted collection time result of the paper archives, and save the predicted collection time result to the electronic backup file; S2: Based on the basic information of each paper archive in the electronic archive and the predicted collection time data, determine whether the collection time requirement is met and perform archive management; S3: Acquire access data and query keyword data of the target user, and determine the query content scope of the target user according to the access data and query keyword data of the target user; S4: using the bag-of-words model to convert the summary content of each file in the database into a summary vector, generating a query value for each file according to the summary vector, and filtering out a preset number of files as query results according to the key type of the target user; S5: Based on the query content, the target user is displayed and an option for the user to log in again for the query is displayed.

2. A smart archive management method based on big data according to claim 1, characterized in that: The method of obtaining relevant data of paper archives and electronic backup files, and using a collection time detection model based on the relevant data to obtain a predicted collection time result of the paper archives, and saving the predicted collection time result in the electronic backup file includes: collecting the collection time, paper material, pen and ink quality, storage environment data and software recognition success rate data of paper archives in different batches, and performing data cleaning to remove outliers and fill in missing values ​​to obtain pre-processed data, and inputting the pre-processed data into the collection time detection model to obtain the predicted collection time of each batch of paper archives; wherein the collection time detection model is a Transformer model trained using historical data and tuned using a cross-validation method; the electronic backup file contains basic information of paper archives that are not included in the electronic archive and the predicted collection time data.

3. A smart archive management method based on big data according to claim 2, characterized in that: Based on the basic information of each paper archive in the electronic backup file and the predicted collection time data, it is judged whether the collection time requirement is met and the archive management is performed, including: judging whether each batch of paper archives will meet the collection time requirement, if the predicted time data of the paper archive is within the first preset time of the current time, the paper archive is added to the archive update sequence; and according to the protection of each paper archive in the archive update sequence, the electronic backup file is updated or entered into the electronic archive.

4. A smart archive management method based on big data according to claim 1, characterized in that: The method of obtaining the target user's access data and query keyword data, and determining the target user's query content range based on the target user's access data and query keyword data, includes: obtaining the target user's query keyword data, and using word embedding technology to map the keyword data to obtain keyword vectors, determining a query range adjustment value based on each keyword vector and the target user's access data; and obtaining the target user's query content range based on the query range adjustment value.

5. A smart archive management method based on big data according to claim 4, characterized in that: The step of determining the query range adjustment value according to each keyword vector and the target user's access data includes: obtaining the historical query times of each keyword in the target user's access data and each keyword vector, and using a query range adjustment formula to obtain the query range adjustment value, wherein the adjustment formula is: ; in is the adjustment value of the query range, is the total number of search keywords, For the query keyword keyword vectors, For keywords The number of historical queries, It is the total number of occurrences of all keywords in historical query records.

6. A smart archive management method based on big data according to claim 4, characterized in that: The step of obtaining the query content range of the target user according to the query range adjustment value includes: obtaining the maximum number of accessible archives of the target user according to the key type input by the target user, and obtaining the query content range of the target user according to the maximum number of accessible archives using an upper limit calculation formula and a lower limit calculation formula, wherein the upper limit calculation formula is: ; In the formula, is the adjustment value of the query range, is the total number of query keywords, T is the maximum number of accessible archives for the target user; The lower limit calculation formula is: .

7. A smart archive management method based on big data according to claim 1, characterized in that: The method uses the bag-of-words model to convert the summary content of each file in the database into a summary vector, generates a query value for each file according to the summary vector, and selects a preset number of files as query results according to the key type of the target user, including: obtaining the key type of the target user when logging in; If it is the first key, the summary content of each file in the database and the summary content in the electronic backup file are converted into a summary vector using the bag-of-words model, and a query value is generated for each file according to the summary vector. According to the query content range and query value of the target user, the files whose query values ​​are within the query content range and greater than or equal to the preset threshold are selected as the query results; If it is the second key, the bag-of-words model is used to convert the summary content of each file in the database into a summary vector, and a query value is generated for each file based on the summary vector. The query content range of the target user is adjusted using the security adjustment formula to obtain an adjusted query content range. Based on the adjusted query content range and the query value, the files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results.

8. A smart archive management method based on big data according to claim 7, characterized in that: If it is the first key, the bag-of-words model is used to convert the summary content of each file in the database and the summary content in the electronic backup file into a summary vector, and a query value is generated for each file according to the summary vector. According to the query content range and query value of the target user, the files whose query values ​​are within the query content range and are greater than or equal to a preset threshold are selected as the query result, including: the first key and the second key are the login keys of the target user, the first key is the standard login key; the second key is the login key set by the target user, and the query value of each file is obtained through the query value calculation formula, wherein the query value calculation formula is: ; In the formula is the query value, is the summary vector corresponding to the summary content of the archive, The total number of words in the abstract content in the archive.

9. A smart archive management method based on big data according to claim 7, characterized in that: If the second key is used, the summary content of each file in the database is converted into a summary vector using a bag of words model, and a query value is generated for each file according to the summary vector, and the query content range of the target user is adjusted using a security adjustment formula to obtain an adjusted query content range, and according to the adjusted query content range and the query value, the files whose query values ​​are within the adjusted query content range and are greater than or equal to a preset threshold are selected as query results, including: adjusting the adjustment value of the query range using a security adjustment formula to obtain a second adjustment value, and obtaining the adjusted query content range according to the second adjustment value, wherein the security adjustment formula is: ; In the formula, is the second adjustment value and Greater than or equal to 0, is the adjustment value of the query range, is the adjustment coefficient, The number of times the second key was used to log in after the first key was used to log in most recently. The maximum number of times tolerable to log in using the second key.

10. A smart archive management method based on big data, according to any one of claims 1 to 9, characterized in that: include: The data collection and preprocessing module is responsible for collecting basic information data of paper archives and electronic backup files, including the collection time, paper material, pen and ink quality, and storage environment of the archives, and performs data denoising, outlier processing, and missing value filling cleaning operations; The collection time prediction module trains the preprocessed archive data based on the collection time detection model, predicts the collection time of each batch of paper archives, and saves the prediction results to the electronic backup file; The archive management module manages archives according to the predicted collection time and archive protection, and determines whether the collection time requirement is met. If the conditions are met, the archive is added to the archive update sequence, and then the electronic backup file is updated or entered into the electronic archive; The user access data analysis module is responsible for obtaining and analyzing the access data and query keyword data of the target user, converting the keyword into a vector using word embedding technology, and determining the query range adjustment value based on the user's historical access data and keyword vector, so as to provide the user with a reasonable query content range in the subsequent steps; The query authority and display module determines the query content range of the user according to the first key or the second key, selects the files whose query values ​​are within the query content range and are greater than or equal to a preset threshold as the query results, and displays them.

Images