Intelligent enterprise archive informatization management system based on big data

Abstract

The invention provides a smart enterprise archive informatization management system based on big data, and belongs to the field of archive information intelligent management systems. The system comprises an information acquisition module; a first information analysis module; an information storage module; a second information analysis module; an application module; according to the system, firstly, user information is analyzed through the big data technology, and the classification conditions of the user information with different marks in all dimensions are obtained; then, a vector formed by multiple dimensions of the user information is analyzed through a support vector machine, and the difference of the overall dimensions of the user information is considered; and finally, the individual difference of the user information in the same classification of the support vector machine is further considered through the LSTM model, so that the prediction result is more personalized and accurate when the demand and the active trend of the user are predicted through the system.

Description

Technical Field

[0001] The present invention relates to the field of intelligent management systems for archive information, and in particular to a smart enterprise archive information management system based on big data. Background Art

[0002] With the introduction of the concept of "Internet +", various Internet information technologies such as cloud computing, big data, and the Internet of Things have penetrated into various fields of society, greatly changing people's lifestyles and work patterns; in corporate management, the advancement of informatization has also provided many conveniences for production, operations, and decision-making; however, in the field of corporate archives management, although it has gradually moved towards computerized management, there are still many problems that need to be solved.

[0003] At present, the archival information of most enterprises, especially those involving important business, legal documents, financial data, etc., still relies on paper documents for preservation and management; this traditional management model not only takes up a lot of physical space, but also easily leads to risks such as loss and damage of archives; the management and retrieval efficiency of paper archives is low. When dealing with a large number of archives, enterprises often face problems such as difficulty in finding and unclear classification, which seriously affects the information flow speed and decision-making efficiency of the enterprise; in addition, with the expansion of the business scale of enterprises, the amount of archival information has increased dramatically, and traditional manual management methods are difficult to cope with efficient information retrieval and updating needs; therefore, there is an urgent need for an electronic archive management system based on modern information technology to improve the management efficiency, security and intelligence level of enterprise archives.

[0004] In this context, the enterprise electronic archive information intelligent management system came into being. Through advanced information technology, it solves the problems existing in the existing archive management methods, realizes the digital and intelligent management of archives, and improves the security and management efficiency of archive storage. The introduction of this system will bring fundamental changes to the enterprise's archive management, help enterprises realize information transformation, and improve overall operational efficiency.

[0005] Big data technology is used to process enterprise information and generate corresponding search reports to facilitate the retrieval and management of enterprise information. This technology is already relatively mature. Big data technology can not only analyze user information in multiple different dimensions, but also analyze user data in the overall dimension. However, what is obtained through big data technology is usually the overall user behavior of the industry, ignoring the differences between individuals. For individuals, the results predicted by big data technology are often far from the actual results. Therefore, a prediction method that can reflect the differences between personalities is needed to accurately predict user needs. Summary of the invention

[0006] The purpose of the present invention is to overcome the above problems existing in the prior art and greatly improve its technical effect on the basis of the original technology; the present invention provides a smart enterprise archive information management system based on big data, the system includes:

[0007] Information acquisition module S100; first information analysis module S200; information storage module S300; second information analysis module S400; application module S500.

[0008] The information acquisition module S100 is connected to the first information analysis module S200, and is used to acquire information and send the acquired information to the first information analysis module S200; the information acquisition module S100 is composed of a passive information acquisition unit and an active information collection unit; the passive information acquisition unit is used to passively acquire information sent by users and information input by employees, and the active information collection unit is used to actively collect browsing information of active users through various network channels.

[0009] The present invention acquires user information by designing two different ports; actively collects user's active information by connecting various network channels through the API port, and passively receives information sent by users and information input by employees through the SMTP port; when acquiring information through different ports, the corresponding user information and employee information will obtain the mark of the corresponding port.

[0010] The first information analysis module S200 is connected to the information acquisition module S100 and the information storage module S300; the first information analysis module S200 includes passively acquired user information analysis and actively acquired user information analysis; the passively acquired user information and the actively acquired user information are analyzed respectively by big data technology, and a passively acquired user information analysis report and an actively acquired user information analysis report are generated, and the analysis report includes the classification of user information in different dimensions; and the generated passively acquired user information analysis report and the actively acquired user information analysis report are sent to the information storage module S300.

[0011] The steps of analyzing passively acquired user information and actively acquired user information through big data technology are: data preprocessing, dimensional feature extraction and data classification; generating easy-to-retrieve passively acquired user information analysis reports and actively acquired user information analysis reports through big data technology, and sending the generated user information analysis reports to the information storage module S300.

[0012] The information storage module S300 is connected to the first information analysis module S200 and the second information analysis module S400; the information storage module S300 includes a passive user information storage area and an active user information storage area; the passive user information storage area is used to store passively obtained user information analysis reports, and the active user information storage area is used to store actively obtained user information analysis reports.

[0013] After obtaining the information transmitted by the first information analysis module S200, first identify through the port mark whether the obtained information is a passively obtained user information analysis report or an actively obtained user information analysis report; according to the mark, the passively obtained user information analysis report is stored in the passive user information storage area, and the actively obtained user information analysis report is stored in the active user information storage area.

[0014] The second information analysis module S400 is connected to the information storage module S300 and the application module S500; the second information analysis module S400 retrieves the user information in the information storage module S300 to obtain the classification of the user information in different dimensions, maps the classification data of different dimensions into a multidimensional space through heterogeneous data integration technology, and generates multidimensional space coordinate information corresponding to the user information; classifies the multidimensional space coordinate information through a support vector machine to obtain a support vector machine model; according to the classification result of the support vector machine model, extracts all user information of each same classification, performs hierarchical assignment on each dimension data of each user information of the same classification, and maps the assigned data of different dimensions after the assignment to the corresponding multidimensional space; regards the multidimensional space data before the assignment as X, and regards the multidimensional space data after the assignment as Y, where X={x1, x2, …x i , …x n}, where x i represents the data corresponding to the i-th dimension, and n represents a total of n dimensions; similarly, Y = {y1, y2, ...y i ,…y n}, where y i represents the data corresponding to the i-th dimension after the value is assigned, and n represents a total of n dimensions; X and Y obtained from each classification in the support vector machine model are used as the input and output of the long short-term memory neural network LSTM, respectively, the LSTM is trained, and the trained LSTM model is obtained.

[0015] Specifically, the second information analysis module S400 obtains the classification of the same labeled user information in different dimensions by respectively retrieving the user information in the passive user information storage area and the active user information storage area in the information storage module S300; the heterogeneous data integration technology is a method of integrating different types of dimensional data into a multidimensional space, and the data standardization and normalization technology in the heterogeneous data integration technology is used to integrate the acquired user information on multiple dimensional classifications into a multidimensional space to generate multidimensional space coordinate information corresponding to the user information, and the support vector machine technology is used to classify the coordinate information in the multidimensional space to generate a support vector machine model; the support vector machine technology is an algorithm commonly used for high-order spatial data classification, which classifies multidimensional data by finding a hyperplane that can maximize the classification interval.

[0016] Specifically, each user information of the same category in the support vector machine model is extracted, and each dimensional data of each user information is hierarchically assigned; the hierarchical assignment includes interval assignment, classification assignment and result assignment; the assigned data of different dimensions after assignment are mapped to the corresponding multidimensional space, and the multidimensional space data obtained after assignment is generated; the long short-term memory neural network LSTM is an improved recursive neural network. LSTM can effectively capture the long-term dependency in time series data by designing complex memory units and gating mechanisms, so that the prediction results are more accurate.

[0017] Specifically, X and Y obtained for each classification in the support vector machine model are used as the input and output of the long short-term memory neural network LSTM, and the LSTM is trained to obtain a trained LSTM model for each classification data in the support vector machine model. That is, if there are A classifications in the same support vector machine model, then A trained LSTM models will be obtained in the end.

[0018] The application module S500 is connected to the second information analysis module S400 and the information acquisition module S100. The application module S500 obtains user information through the client and sends the user information to the information acquisition module S100. Then, the user information is classified into different dimensions through the big data technology of the first information analysis module S200, and the classified user information is saved in the information storage module S300. At the same time, the user information classified into different dimensions is mapped into a multidimensional space to form a vector in the multidimensional space. The obtained vector in the multidimensional space is respectively brought into the support vector machine model and the trained LSTM model in the second information analysis module S400, and the user's needs are judged according to the results.

[0019] The client is connected to the user's network device, obtains user information through the user's browsing information, and sends the user information to the information acquisition module S100. The information acquisition module S100 will determine whether the user information is passively acquired or actively collected based on the format of the user information, and generate a tag for the corresponding user information; then, the big data technology is used to process the user information, generate a user information analysis report with tag information, and save the user information analysis report to the corresponding area of ​​the information storage module S300 based on the tag information; at the same time, the classification information of different dimensions in the user information analysis report is sent to the second data analysis module, and the second data analysis module maps the obtained classification information of different dimensions of the user into a multidimensional space to form a vector in the multidimensional space, and respectively brings the obtained vector of the multidimensional space into the support vector machine model and the trained LSTM model in the second information analysis module S400, and determines the user's activity, demand and shopping possibility based on the results output by the LSTM model.

[0020] The beneficial effects of the present invention are:

[0021] The present invention provides a smart enterprise archive information management system based on big data; the system has the following advantages:

[0022] 1. The system divides user information into passively obtained user information and actively obtained user information, and stores and analyzes them separately, so that the analysis results will be more targeted.

[0023] 2. When analyzing user information, the system first classifies the user information in various dimensions through big data technology, then extracts the various dimensional data of each user information in the big data information classification to form a multidimensional space data, and classifies all multidimensional space coordinate information with the same label through support vector machine to obtain a support vector machine model; then extracts the various dimensional data of user information of the same classification in the support vector machine model for hierarchical assignment; the vectors composed of user information before assignment and user information after assignment are used as input and output of LSTM to train LSTM, and prediction is made through the trained LSTM model; the system not only considers the overall dimensional gap of user information through support vector machine, but also further considers the individual differences of user information in the same classification of support vector machine through LSTM model. Therefore, when the system predicts user needs and trends, the prediction results will be more personalized and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of a smart enterprise archive information management system based on big data of the present invention. DETAILED DESCRIPTION

[0025] The specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings; it should be understood that the specific embodiments given here are only used to illustrate and explain the present invention and cannot be used to limit the present invention.

[0026] It should be noted that many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention may also have other implementations and variations thereof. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0027] like Figure 1 As shown, a schematic diagram of a smart enterprise archive information management system based on big data according to an embodiment of the present invention includes: an information acquisition module S100; a first information analysis module S200; an information storage module S300; a second information analysis module S400; and an application module S500; wherein the information acquisition module S100 is composed of a passive information acquisition unit S101 and an active information collection unit S102, and the information storage module S300 is composed of a passive user information storage area S301 and an active user information storage area S302.

[0028] The information acquisition module S100 is connected to the first information analysis module S200, and is used to acquire information and send the acquired information to the first information analysis module S200; the information acquisition module S100 is composed of a passive information acquisition unit S101 and an active information collection unit S102; the passive information acquisition unit S101 is used to passively acquire information sent by users and information input by employees, and the active information collection unit S102 is used to actively collect browsing information of active users through various network channels.

[0029] Specifically, user information is obtained by designing two different ports; the user's active information is actively collected through various network channels connected through the API port, and the information sent by the user and the information entered by the employee are passively received through the SMTP port; when information is obtained through different ports, the corresponding user information and employee information will obtain the mark of the corresponding port; that is, the user's active information actively collected through the API port will generate an API port identifier, and the information sent by the user and the information entered by the employee passively received through the SMTP port will generate an SMTP port identifier, and the identifier can be set arbitrarily, as long as the relevant equipment and algorithms are easy to distinguish; therefore, the port of the information acquisition module S100 is composed of two ports, an API port for actively collecting the user's active information, in preparation for subsequent understanding of the user's activity and prediction of the user's needs; and an SMTP port for passively receiving the information sent by the user and the information entered by the employee, in preparation for subsequent understanding of the user's needs and the employee's work status.

[0030] The first information analysis module S200 is connected to the information acquisition module S100 and the information storage module S300; the first information analysis module S200 analyzes the passively obtained user information and the actively obtained user information; the passively obtained user information and the actively obtained user information are analyzed respectively by big data technology, and a passively obtained user information analysis report and an actively obtained user information analysis report are generated, and the analysis report includes the classification of user information in different dimensions; and the generated passively obtained user information analysis report and the actively obtained user information analysis report are sent to the information storage module S300.

[0031] The steps of analyzing passively acquired user information and actively acquired user information through big data technology are: data preprocessing, dimensional feature extraction and data classification, etc.; generating easy-to-retrieve passively acquired user information analysis reports and actively acquired user information analysis reports through big data technology, and sending the generated user information analysis reports to the information storage module S300.

[0032] The user information analysis report generated by big data technology can obtain user information and enterprise-related information related to the entry in the form of retrieval.

[0033] It should be pointed out that when big data technology analyzes relevant user information, it uses cleaning and automatic filtering functions to filter out and clean illegal information and information that does not comply with corporate regulations.

[0034] It should be pointed out that the first information analysis module S200 analyzes the information pages passively obtained by employees at the same time as analyzing the passively obtained user information, generates relevant entries of the employee input information and the relationship between the information input by employees and users based on big data technology, and generates an analysis report of the information input by employees together with the passively obtained user information, and transmits the analysis report to the information storage module S300.

[0035] The information storage module S300 is connected to the first information analysis module S200 and the second information analysis module S400; the information storage module S300 is divided into a passive user information storage area and an active user information storage area; the passive user information storage area is used to store passively obtained user information analysis reports, and the active user information storage area is used to store actively obtained user information analysis reports.

[0036] After obtaining the information transmitted by the first information analysis module S200, the information storage module S300 first identifies through the port mark whether the obtained information is a passively obtained user information analysis report or an actively obtained user information analysis report; according to the mark, the passively obtained user information analysis report is stored in the passive user information storage area, and the actively obtained user information analysis report is stored in the active user information storage area.

[0037] The second information analysis module S400 is connected to the information storage module S300 and the application module S500; the second information analysis module S400 retrieves the user information in the information storage module S300, obtains the classification of the user information in different dimensions, maps the classified data of different dimensions into a multidimensional space through heterogeneous data integration technology, and generates multidimensional space coordinate information corresponding to the user information; classifies the multidimensional space coordinate information through a support vector machine to obtain a support vector machine model.

[0038] By respectively retrieving user information in the passive user information storage area and the active user information storage area in the information storage module S300, the classification of the same labeled user information in different dimensions is obtained, and a multidimensional space vector is formed through heterogeneous data integration technology according to the classification data of each user information in different dimensions; heterogeneous data integration technology is a method of integrating different types of dimensional data into a multidimensional space, and the data of multiple dimensional classifications of the obtained user information are integrated into a multidimensional space through the data standardization and normalization technology in the heterogeneous data integration technology to generate multidimensional space coordinate information corresponding to the user information, and then the coordinate information in the multidimensional space is classified through the support vector machine technology to generate a support vector machine model; support vector machine technology is an algorithm commonly used for high-order spatial data classification, which classifies multidimensional data by finding a hyperplane that can maximize the classification interval.

[0039] It should be pointed out that the present invention obtains information through two ports, each port has a label; therefore, a total of two types of labeled user information are obtained in different dimensions; when the coordinate information of different labels in the multidimensional space is classified by the support vector machine, a support vector machine model will be generated for each type of labeled user information, and the present invention generates two different support vector machine models.

[0040] According to the classification results of the support vector machine model, all user information of each same category is extracted, and each dimension data of each user information of the same category is hierarchically assigned, and the assigned data of different dimensions after the assignment is mapped to the corresponding multidimensional space; the multidimensional space data before the assignment is regarded as X, and the multidimensional space data after the assignment is regarded as Y, where X = {x1, x2, ...x i , …x n}, where xi represents the data corresponding to the i-th dimension, and n represents a total of n dimensions; similarly, Y = {y1, y2, ...y i ,…y n}, where y i It represents the data corresponding to the i-th dimension after the value is assigned, and n represents that there are n dimensions in total.

[0041] The hierarchical assignment of each dimension data of each user information of the same category includes: extracting each user information of the same category in the support vector machine model, and hierarchical assignment of each dimension data of each user information; the hierarchical assignment includes interval assignment, classification assignment and result assignment; mapping the assigned data of different dimensions after the assignment to the corresponding multidimensional space, and generating the multidimensional space data after the assignment; the long short-term memory neural network LSTM is an improved recursive neural network. LSTM can effectively capture the long-term dependency in the time series data by designing complex memory units and gating mechanisms, so that the prediction results are more accurate.

[0042] Preferably, for example: an interval is assigned to the time corresponding to the time when the user information is obtained, and a day is regarded as 24 hours, 0:00 to 0:30 can be assigned 1, 0:30 to 1:00 can be assigned 2, and so on, until the entire 24-hour time period of the day is assigned; the consumption level of the obtained user information is classified and assigned, and monthly consumption below 500 is regarded as a low consumption level and assigned a value of 1; consumption between 501 and 2000 is regarded as an equal consumption level and assigned a value of 2; consumption above 2001 is regarded as a high consumption level and assigned a value of 3; the result of whether the user information is consumed is assigned, if the user consumes, it is assigned a value of 1, if the user does not consume, it is assigned a value of 0.

[0043] The X and Y obtained from each classification in the support vector machine model are used as the input and output of the long short-term memory neural network LSTM, the LSTM is trained, and a trained LSTM model is obtained.

[0044] The X and Y obtained for each classification in the support vector machine model are used as the input and output of the long short-term memory neural network LSTM respectively, and the LSTM is trained to obtain the trained LSTM model corresponding to each classification data in the support vector machine model. That is, if there are A classifications in the same support vector machine model, then A trained LSTM models will be obtained in the end.

[0045] The application module S500 is connected to the second information analysis module S400 and the information acquisition module S100. The application module S500 obtains user information through the client and sends the user information to the information acquisition module S100. Then, the user information is classified into different dimensions through the big data technology of the first information analysis module S200, and the classified user information is saved in the information storage module S300. At the same time, the user information classified into different dimensions is mapped into a multidimensional space to form a vector in the multidimensional space. The obtained vector in the multidimensional space is respectively brought into the support vector machine model and the trained LSTM model in the second information analysis module S400, and the user's needs are judged according to the results.

[0046] The client is connected to the user's network device, obtains user information through the user's browsing information, and sends the user information to the information acquisition module S100. The information acquisition module S100 determines whether the user information is passively acquired or actively collected according to the format of the user information, and generates a tag for the corresponding user information; then, the big data technology processes the user information, generates a user information analysis report with tag information, and saves the user information analysis report to the corresponding area of ​​the information storage module S300 according to the tag information; at the same time, the classification information of different dimensions in the user information analysis report is sent to the second data analysis module, The second data analysis module maps the obtained classification information of users in different dimensions into a multidimensional space to form a vector in the multidimensional space, and respectively brings the obtained vector in the multidimensional space into the support vector machine model and the trained LSTM model in the second information analysis module S400, and determines the user's activity, demand and shopping possibility according to the output results of the LSTM model; the above process needs to be carried out separately from the establishment of the support vector machine model and the LSTM model. Therefore, the data of this process needs to be transmitted and analyzed separately from the data obtained by the information acquisition module S100, and the above technical features can be implemented through a multi-threaded system.

[0047] The above-mentioned embodiments of the present invention are implemented by arranging servers, routers, switches, sensors, Internet of Things devices, visualization tools, and intelligent algorithms in different modules and unit areas.

Claims

1. A smart enterprise archive information management system based on big data, characterized in that: The system comprises: Information acquisition module S100; first information analysis module S200; information storage module S300; second information analysis module S400; application module S500; The information acquisition module S100 is connected to the first information analysis module S200, and is used to acquire information and send the acquired information to the first information analysis module S200; the information acquisition module S100 is composed of a passive information acquisition unit and an active information acquisition unit; the passive information acquisition unit is used to passively acquire information sent by users and information input by employees, and the active information acquisition unit is used to actively acquire browsing information of active users through various network channels; The first information analysis module S200 is connected to the information acquisition module S100 and the information storage module S300; the first information analysis module S200 includes passively obtained user information analysis and actively obtained user information analysis; the passively obtained user information and the actively obtained user information are analyzed by big data technology respectively, and a passively obtained user information analysis report and an actively obtained user information analysis report are generated, and the analysis report includes classification of user information in different dimensions; and the generated passively obtained user information analysis report and the actively obtained user information analysis report are sent to the information storage module S300; The information storage module S300 is connected to the first information analysis module S200 and the second information analysis module S400; the information storage module S300 includes a passive user information storage area and an active user information storage area; the passive user information storage area is used to store passively obtained user information analysis reports, and the active user information storage area is used to store actively obtained user information analysis reports; The second information analysis module S400 is connected to the information storage module S300 and the application module S500; the second information analysis module S400 retrieves the user information in the information storage module S300, obtains the classification of the user information in different dimensions, maps the classified data of different dimensions into a multi-dimensional space through heterogeneous data integration technology, and generates multi-dimensional space coordinate information corresponding to the user information; classifies the multi-dimensional space coordinate information through a support vector machine to obtain a support vector machine model; According to the classification results of the support vector machine model, all user information of each same category is extracted, and each dimension data of each user information of the same category is hierarchically assigned, and the assigned data of different dimensions after the assignment is mapped to the corresponding multidimensional space; the multidimensional space data before the assignment is regarded as X, and the multidimensional space data after the assignment is regarded as Y, where X = {x1, x2, ...x i , …x n }, where x i represents the data corresponding to the i-th dimension, and n represents a total of n dimensions; similarly, Y = {y1, y2, ...y i ,…y n }, where y i It indicates the data corresponding to the i-th dimension after the value is assigned, and n indicates that there are n dimensions in total; The X and Y obtained from each classification in the support vector machine model are used as the input and output of the long short-term memory neural network LSTM, the LSTM is trained, and the trained LSTM model is obtained; The application module S500 is connected to the second information analysis module S400 and the information acquisition module S100. The application module S500 obtains user information through the client and sends the user information to the information acquisition module S100. Then, the user information is classified in different dimensions through the big data technology of the first information analysis module S200, and the classified user information is saved in the information storage module S300. At the same time, the user information after classification in different dimensions is mapped into a multidimensional space to form a vector in the multidimensional space. The obtained vector in the multidimensional space is respectively brought into the support vector machine model and the trained LSTM model in the second information analysis module S400, and the user's needs are judged according to the results.

2. According to the big data-based smart enterprise archive information management system of claim 1, it is characterized in that: The information acquisition module S100 includes: acquiring user information by designing two different ports; actively collecting user's active information by connecting to various network channels through an API port, and passively receiving information sent by users and information input by employees through an SMTP port; when acquiring information through different ports, the corresponding user information and employee information will obtain the mark of the corresponding port.

3. According to the big data-based smart enterprise archive information management system of claim 1, it is characterized in that: The first information analysis module S200 includes: the steps of analyzing passively obtained user information and actively obtained user information through big data technology: data preprocessing, dimensional feature extraction and data classification; generating easy-to-retrieve passively obtained user information analysis reports and actively obtained user information analysis reports through big data technology, and sending the generated user information analysis reports to the information storage module S300.

4. According to the big data-based smart enterprise archive information management system of claim 1, it is characterized in that: The information storage module S300 includes: after obtaining the information transmitted by the first information analysis module S200, first identify through the port mark whether the obtained information is a passively obtained user information analysis report or an actively obtained user information analysis report; according to the mark, the passively obtained user information analysis report is stored in the passive user information storage area, and the actively obtained user information analysis report is stored in the active user information storage area.

5. According to the big data-based smart enterprise archive information management system of claim 1, it is characterized in that: The second information analysis module S400 includes: the second information analysis module S400 obtains the classification of the same labeled user information in different dimensions by retrieving the user information in the passive user information storage area and the active user information storage area in the information storage module S300 respectively; the heterogeneous data integration technology is a method of integrating different types of dimensional data into a multidimensional space, and the data on multiple dimensional classifications of the obtained user information are integrated into a multidimensional space through the data standardization and normalization technology in the heterogeneous data integration technology to generate multidimensional space coordinate information corresponding to the user information, and the coordinate information on the multidimensional space is classified through the support vector machine technology to generate a support vector machine model; the support vector machine technology is an algorithm commonly used for high-order spatial data classification, which classifies multidimensional data by finding a hyperplane that can maximize the classification interval.

6. According to the big data-based smart enterprise archive information management system of claim 1, it is characterized in that: The hierarchical assignment of each dimension data of each user information of the same category includes: extracting each user information of the same category in the support vector machine model, and hierarchical assignment of each dimension data of each user information; the hierarchical assignment includes interval assignment, classification assignment and result assignment; mapping the assigned data of different dimensions after assignment to the corresponding multidimensional space, generating multidimensional space data after assignment; the long short-term memory neural network LSTM is an improved recursive neural network. LSTM can effectively capture the long-term dependency in time series data by designing complex memory units and gating mechanisms, so that the prediction results are more accurate.

7. The big data-based smart enterprise archive information management system according to claim 1 is characterized in that: The trained LSTM model includes: using X and Y obtained from each classification in the support vector machine model as input and output of the long short-term memory neural network LSTM, respectively, training the LSTM, and obtaining a trained LSTM model corresponding to each classification data in the support vector machine model, that is, if there are A classifications in the same support vector machine model, then A trained LSTM models are finally obtained.

8. The big data-based smart enterprise archive information management system according to claim 1 is characterized in that: The application module S500 includes: a client is connected to the user's network device, obtains user information through the user's browsing information, and sends the user information to the information acquisition module S100. The information acquisition module S100 determines whether the user information is passively acquired or actively collected according to the format of the user information, and generates a tag for the corresponding user information; then, the big data technology processes the user information, generates a user information analysis report with tag information, and saves the user information analysis report to the corresponding area of ​​the information storage module S300 according to the tag information; at the same time, the classification information of different dimensions in the user information analysis report is sent to the second data analysis module, and the second data analysis module maps the obtained classification information of different dimensions of the user into a multidimensional space to form a vector in the multidimensional space, and respectively brings the obtained vector of the multidimensional space into the support vector machine model and the trained LSTM model in the second information analysis module S400, and determines the user's activity, demand and shopping possibility according to the results output by the LSTM model.

Images