A method for intelligent analysis of massive information based on big data

Abstract

The present invention belongs to the field of information analysis management, and relates to a method for intelligent analysis of massive information based on big data. The method integrates keyword tags, subject tags, time tags and space tags of data items newly introduced on the day, generates index tags thereof and uses them as the pointing basis for storing in a target database, thereby greatly improving the data management quality of the target database, collecting the retrieval data of each archived data item in the target database as of a preset time point on the day, identifying each high-frequency data item and each low-frequency data item in the target database, optimizing the index tags and storage locations of the high-frequency data items, transferring the low-frequency data items to a special low-frequency data library, identifying the oversaturation state of the special low-frequency data library and further performing data cleaning operations, thereby not only improving the retrieval efficiency and storage resource utilization of the target database, but also reducing storage costs and enhancing data management capabilities.

Description

Technical Field

[0001] The present invention belongs to the field of information analysis management and relates to a method for intelligent analysis of massive information based on big data. Background Art

[0002] In the context of the digital age, the amount of data is growing exponentially, the scale and complexity of information are increasing, and all walks of life continue to generate and store massive amounts of data in various databases. Faced with this trend, efficient information analysis and management of data stored in databases has become an urgent need.

[0003] In the prior art, there are also some solutions related to database data storage information analysis, for example, the database log storage method, database log analysis method and device with Chinese patent publication number CN114817245A, which obtains the real-time log information of the database and performs a first write operation to write the real-time log information into a system log table pre-established in the database. Although it can effectively guide the management of database data storage information, it still has the limitations commonly found in the prior art, specifically: 1. The prior art usually generates index tags for imported data items to facilitate their database storage, but the generation of data item index tags depends on the extraction of keyword frequency from its text content, ignoring the overall distribution of the target database and the semantic similarity between keywords, which may lead to some common but non-discriminatory words being mistakenly selected as keywords, thereby making the keyword tag retrieval unreliable. In addition, relying solely on keyword extraction to construct index tags can easily make the index tags unreliable, unreferenceable and uneffective.

[0004] 2. The prior art lacks an optimization mechanism for the database storage location during the subsequent retrieval process of data items, and is unable to timely integrate the actual retrieval information of data items to optimize the management of internal storage in the database. Summary of the invention

[0005] In view of this, in order to solve the problems raised in the above background technology, a method for intelligent analysis of massive information based on big data is proposed.

[0006] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides a method for intelligent analysis of massive information based on big data, including: S1. Importing each data item newly introduced on the day into the target database, parsing the text content of each data item and generating its corresponding index label.

[0007] S2. Based on the index tags, each newly introduced data item of the day is stored in the target database and its corresponding associated sub-databases.

[0008] S3. Perform a target database traversal operation at a preset time point on the same day, collect the retrieval data of each archived data item in the target database as of the preset time point on the same day, including the number of retrievals, the date of each retrieval, the input retrieval content and the corresponding call associated sub-library, and filter out each high-frequency data item and each low-frequency data item in the target database.

[0009] S4. Optimize the index label and storage location of each high-frequency data item in the target database.

[0010] S5. Each low-frequency data item in the target database is transferred to a special low-frequency data database. If it is identified that the special low-frequency data database is oversaturated after the transfer process, data cleaning is performed on the special low-frequency data database.

[0011] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention determines the importance evaluation coefficient of each word in the data item text word set by carefully considering the word frequency within the word compared with its data item and the inverse document frequency within the word compared with the target database, and combines the repeated semantic keyword removal operation to help the data item generate refined and accurate keyword tags, thereby achieving representativeness, effectiveness and reliability of data item keyword extraction.

[0012] (2) The present invention integrates the keyword tags, subject tags, time tags and space tags of data items to generate their index tags and uses them as the reference for storing them in the target database. This not only effectively improves the efficiency and accuracy of data retrieval in the target database, but also facilitates the subsequent data management and maintenance of the target database.

[0013] (3) The present invention identifies each high-frequency data item and each low-frequency data item in the target database by retrieving data of each archived data item in the target database as of a preset time point on the same day, thereby ensuring that the data analysis of the subsequent processing of the data items has sufficient security and depth, which not only improves the accuracy and pertinence of the data analysis, but also promotes the rational allocation of resources and the maximization of data value.

[0014] (4) The present invention implements different processing strategies for high-frequency data items and low-frequency data items in the target database, thereby not only improving the retrieval efficiency and storage resource utilization of the target database, but also reducing storage costs and enhancing data management capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0016] Figure 1The present invention is a flowchart of the steps for implementing the method. 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] See also Figure 1 As shown, the present invention provides a method for intelligent analysis of massive information based on big data, including: S1. importing each data item newly introduced on the day into the target database, parsing the text content of each data item and generating its corresponding index label.

[0019] Specifically, the method of parsing the text content of each data item and generating its corresponding index tag includes: removing the stop words involved in the text content of each data item according to a predefined stop word list stored in the WEB cloud, sorting the text word set of each data item after word segmentation, retrieving the word frequency of each word in the text word set compared with its data item and the inverse document frequency of the word in the target database, taking the product of the two as the importance evaluation coefficient, screening out the words whose importance evaluation coefficient is greater than or equal to the preset importance evaluation coefficient reaching threshold value, recording them as keywords, thereby obtaining the keywords in the text word set of each data item.

[0020] Determine the part-of-speech category of each keyword in the text word set of each data item, compare the semantic similarity between keywords under the same part-of-speech category in the text word set of each data item, match semantically similar keyword pairs under the same part-of-speech category, compare the values ​​of importance assessment coefficients between keywords in the semantically similar keyword pairs, eliminate the keywords corresponding to the one with the smaller value of importance assessment coefficient in the semantically similar keyword pairs, and retain the keywords belonging to each part of speech in the text word set of each data item as keyword labels for the corresponding data items.

[0021] It should be noted that the semantic similarity between the keywords under the same part-of-speech category in the above-mentioned text word sets of each data item is obtained by substituting the keywords under the same part-of-speech category in the text word sets of each data item into the word embedding model and converting them into vectors, and obtaining the semantic similarity between the keywords under the same part-of-speech category in the text word sets of each data item through the cosine similarity calculation formula. If the semantic similarity between a pair of keywords is greater than or equal to the semantic similarity threshold corresponding to the semantic similarity state, then the pair of keywords is judged to be a semantically similar keyword pair.

[0022] The embodiment of the present invention determines the importance assessment coefficient of each word in the data item text word set by carefully considering the word frequency within the word compared with its data item and the inverse document frequency within the word compared with the target database, and combines the repeated semantic keyword removal operation to help the data item generate refined and accurate keyword tags, thereby achieving representativeness, effectiveness and reliability of data item keyword extraction.

[0023] According to the vocabulary list of each text topic stored in the WEB cloud, the word distribution probability of each data item text word set relative to each text topic is retrieved, and the text topic corresponding to the largest word distribution probability is selected as the topic label of the data item, thereby obtaining the topic label of each data item.

[0024] It should be noted that the word distribution probability of each data item text word set relative to each text topic refers to the ratio of the cumulative number of words involved in each data item text word set relative to each text topic to the total number of words in its data item text word set.

[0025] The time attribute words and the spatial location attribute words in the text content of each data item are retrieved, and the time label and the space label of each data item are generated in the corresponding preset standard format.

[0026] Integrate keyword tags, topic tags, time tags, and space tags to generate index tags for each data item.

[0027] The embodiment of the present invention integrates the keyword tags, subject tags, time tags and space tags of data items to generate their index tags and uses them as the pointing basis for storing them in the target database, which not only effectively improves the efficiency and accuracy of data retrieval in the target database, but also facilitates subsequent data management and maintenance of the target database.

[0028] Specifically, the word frequency of each word in the retrieval text word set compared with its data item and the inverse document frequency of the word in the retrieval text word set compared with the target database include: counting the occurrence frequency of each word in the text word set in the text content of its data item, performing a ratio analysis with the total number of words in the data item text word set, and obtaining the word frequency of each word in the text word set compared with its data item.

[0029] Count the number of archived data items in the target database whose text content contains a word in the text word set as of the current day, perform ratio analysis on it with the total number of archived data items in the target database as of the current day, take the natural logarithm of the ratio analysis result as the inverse document frequency of the word in the text word set compared with the target database, and then obtain the inverse document frequency of each word in the text word set compared with the target database.

[0030] S2. Based on the index tags, each newly introduced data item of the day is stored in the target database and its corresponding associated sub-databases.

[0031] Specifically, the storing of each newly introduced data item on the day into the target database and its corresponding associated sub-libraries includes: retrieving the standard index tags of each sub-library in the target database, including the keyword standard tags, the subject standard tags, the time standard tags and the space standard tags, comparing them with the index tags of each newly introduced data item on the day, and analyzing the correlation coefficient between the keyword tags of each newly introduced data item on the day and each sub-library in the target database. , topic label correlation coefficient , time label correlation coefficient and spatial label association coefficient , is the number of each data item newly introduced on that day, , is the number of each sub-database in the target database, , according to the formula Calculate the comprehensive correlation index between each newly introduced data item and each sub-database in the target database on that day, where The preset weights correspond to the keyword tag association coefficient, the topic tag association coefficient, the time tag association coefficient, and the space tag association coefficient.

[0032] For example, the above The specific values ​​can be 0.3, 0.3, 0.2, 0.2.

[0033] If the comprehensive correlation index between a certain data item newly introduced on the day and a certain sub-library in the target database is greater than or equal to the preset comprehensive correlation index reaching threshold, the sub-library will be used as the associated sub-library of the data item, and then the associated sub-libraries corresponding to the newly introduced data items on the day will be obtained and the storage operation will be performed accordingly.

[0034] Specifically, the and The specific analysis process includes: representing the keyword labels of each sub-database in the target database and the keyword labels of each data item newly introduced on the day in the form of a set, respectively recorded as .

[0035] The importance evaluation coefficient of each keyword in the keyword tag of each data item newly introduced on the day is extracted, the importance weight of each keyword in the keyword tag of each data item newly introduced on the day is determined, and then the importance weight of the corresponding set element is mapped.

[0036] It should be noted that the specific process of determining the importance weight of each keyword in the keyword tag of each data item newly introduced on the above-mentioned day is: accumulating the importance evaluation coefficient of each keyword in the keyword tag of a data item newly introduced on the day to obtain the overall importance evaluation coefficient of the keyword tag of the data item newly introduced on the day, performing a ratio analysis of the importance evaluation coefficient of each keyword in the keyword tag of the data item newly introduced on the day and its overall importance evaluation coefficient, so as to obtain the importance weight of each keyword in the keyword tag of the data item newly introduced on the day, thereby determining the importance weight of each keyword in the keyword tag of each data item newly introduced on the day.

[0037] By formula Analyze the correlation coefficient between each newly introduced data item and the keyword label of each sub-database in the target database, where They represent the newly introduced The keyword tag set of each data item is the same as the keyword tag set of the target database. The number of elements in the intersection and union of the keyword tag sets of the sub-libraries, Indicates the newly introduced The keyword tag set of each data item is the same as the keyword tag set of the target database. The keyword tag set of the sub-library The importance weights of the intersection elements, is the number of each intersection element, .

[0038] According to the predefined subject words corresponding to the replacement vectors stored in the WEB cloud, the subject labels of each sub-database in the target database and the subject labels of each data item newly introduced on the day are represented in the form of vectors, which are recorded as , according to the formula Analyze the correlation coefficient between each newly introduced data item on that day and the subject label of each sub-database in the target database.

[0039] Specifically, the and The specific analysis process includes: recording the time span corresponding to the time label of each sub-database in the target database and the time span corresponding to the time label of each data item newly introduced on the day as , according to the calculation model Analyze the correlation coefficient between each newly introduced data item and the time label of each sub-database in the target database, where .

[0040] It should be noted that the above Indicates the newly introduced The time tag of each data item corresponds to the time span contained in the target database. The time labels of each sub-database correspond to the time span.

[0041] Similarly, the spatial labels of each sub-database in the target database correspond to the regional categories and the spatial labels of each data item newly introduced on the day correspond to the regional categories. According to the inclusion relationship between the regional categories, the correlation coefficients of the spatial labels of each data item newly introduced on the day and each sub-database in the target database are analyzed.

[0042] S3. Perform a target database traversal operation at a preset time point on the same day, collect the retrieval data of each archived data item in the target database as of the preset time point on the same day, including the number of retrievals, the date of each retrieval, the input retrieval content and the corresponding call associated sub-library, and filter out each high-frequency data item and each low-frequency data item in the target database.

[0043] Specifically, the screening of each high-frequency data item and each low-frequency data item in the target database includes: calculating the time span between each retrieval date of each archived data item in the target database and its last retrieval date as of the preset time point on the current day, and obtaining the reference retrieval period of each archived data item in the target database by average calculation; , combined with the number of retrievals of each archived data item in the target database as of the preset time point on that day , analyze the retrieval frequency of each archived data item in the target database , is the number of each archived data item in the target database, .

[0044] It should be noted that the calculation formula for the retrieval frequency of each archived data item in the above target database is: ,in The number of archived data items in the target database.

[0045] The high-frequency retrieval status of data items stored in the WEB cloud corresponds to the retrieval frequency threshold The retrieval frequency threshold corresponding to the low-frequency retrieval state , will satisfy The archived data items of the condition are regarded as high-frequency data items, which will satisfy The archived data items of the conditions are used as low-frequency data items to filter out the high-frequency data items and low-frequency data items in the target database.

[0046] The embodiment of the present invention identifies each high-frequency data item and each low-frequency data item in the target database by retrieving data of each archived data item in the target database as of a preset time point on the same day, thereby ensuring that data analysis for subsequent processing of the data items has sufficient security and depth, which not only improves the accuracy and pertinence of data analysis, but also promotes the rational allocation of resources and maximization of data value.

[0047] S4. Optimize the index label and storage location of each high-frequency data item in the target database.

[0048] Specifically, the index tag optimization processing is performed on each high-frequency data item in the target database, including: extracting the input search content of each search of each high-frequency data item in the target database as of a preset time point on the same day, finding the newly added words and semantic deviation degree of the input search content of each search of the high-frequency data item relative to the index tag, analyzing the optimization demand coefficient of the index tag of each high-frequency data item in the target database, identifying that the index tag of each high-frequency data item in the target database whose optimization demand coefficient is greater than or equal to a preset optimization demand coefficient warning threshold has optimization demand, and selecting a preset number of valid words from the newly added words of the index tag of the input search content of each search, and adding them to the index tag.

[0049] It should be noted that the calculation formula for the optimization requirement coefficient of each high-frequency data item index label in the above target database is: ,in The target database High-frequency data item The input search content of the search is relative to the index tag The semantic deviation of the newly added words, is the number of each high-frequency data item in the target database, , is the number of each search, , The numbers of the newly added words relative to the index tags for the search content entered. , The target database High-frequency data item The number of new words added to the input search content relative to the index tag of the search, is the number of searches, They are respectively the preset number of reference newly added words and the semantic deviation degree of reference newly added words.

[0050] Specifically, the storage location optimization processing for each high-frequency data item in the target database includes: extracting the index label of each high-frequency data item in the target database after the optimization processing, analyzing the comprehensive correlation index between each high-frequency data item in the target database and each sub-library, and then screening each associated sub-library corresponding to the optimized index label of each high-frequency data item in the target database, comparing with each associated sub-library stored in correspondence with each high-frequency data item, and determining each newly added associated sub-library corresponding to each high-frequency data item.

[0051] The corresponding call associated sub-libraries of each high-frequency data item in the target database as of the preset time point on the day are extracted, and the corresponding search activity of each associated sub-library in which each high-frequency data item in the target database has been stored is evaluated, and the associated sub-libraries whose search activity is greater than or equal to the preset search activity warning threshold are taken as the retained associated sub-libraries.

[0052] The newly added associated sub-libraries and the retained associated sub-libraries corresponding to the high-frequency data items are integrated to obtain the associated sub-libraries stored corresponding to the high-frequency data items after optimization processing.

[0053] S5. Each low-frequency data item in the target database is transferred to a special low-frequency data database. If it is identified that the special low-frequency data database is oversaturated after the transfer process, data cleaning is performed on the special low-frequency data database.

[0054] Specifically, the data cleaning process for the low-frequency data database includes: determining the number of oversaturated data items in the low-frequency data database, extracting the search frequency of each low-frequency data item in the low-frequency data database, combining the import date of each low-frequency data item to the target database, , is the number of each low-frequency data item in the low-frequency data database, , analyze the retrieval value coefficient of each low-frequency data item in the low-frequency data database, and clean up the low-frequency data items whose retrieval value coefficient is ranked after the number of oversaturated data items.

[0055] It should be noted that the calculation formula for the retrieval value coefficient of each low-frequency data item in the above-mentioned low-frequency data database is: ,in It is the first The retrieval frequency of low-frequency data items, is today's date, The preset reference time span.

[0056] The embodiment of the present invention not only improves the retrieval efficiency and storage resource utilization of the target database, but also reduces storage costs and enhances data management capabilities by performing different processing strategies on high-frequency data items and low-frequency data items in the target database.

[0057] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A method for intelligent analysis of massive information based on big data, characterized in that: The method includes: S1. Import each newly introduced data item of the day into the target database, parse the text content of each data item and generate its corresponding index label; S2. According to the index tag, each newly introduced data item of the day is stored in the target database and its corresponding associated sub-database; S3. Perform a target database traversal operation at a preset time point on the same day, collect the search data of each archived data item in the target database as of the preset time point on the same day, including the number of searches, the date of each search, the input search content and the corresponding call associated sub-library, and filter each high-frequency data item and each low-frequency data item in the target database; S4. Optimizing the index label and storage location of each high-frequency data item in the target database; S5. Transferring each low-frequency data item in the target database to a low-frequency data database. If the low-frequency data database is found to be oversaturated after the transfer, data cleaning is performed on the low-frequency data database. The storing of each newly introduced data item on the day into the target database and its corresponding associated sub-libraries includes: retrieving the standard index tags of each sub-library in the target database, including the keyword standard tag, the subject standard tag, the time standard tag and the space standard tag, comparing them with the index tags of each newly introduced data item on the day, and analyzing the correlation coefficient between each newly introduced data item on the day and the keyword tag of each sub-library in the target database , topic label correlation coefficient , time label correlation coefficient and spatial label association coefficient , is the number of each data item newly introduced on that day, , is the number of each sub-database in the target database, , according to the formula Calculate the comprehensive correlation index between each newly introduced data item and each sub-database in the target database on that day, where The preset weights correspond to the keyword tag association coefficient, the topic tag association coefficient, the time tag association coefficient, and the space tag association coefficient respectively; If the comprehensive correlation index between a certain data item newly introduced on the day and a certain sub-database in the target database is greater than or equal to the preset comprehensive correlation index reaching threshold, the sub-database is used as the associated sub-database of the data item, and then the associated sub-databases corresponding to the data items newly introduced on the day are obtained and the storage operation is performed accordingly; Said and The specific analysis process includes: representing the keyword labels of each sub-database in the target database and the keyword labels of each data item newly introduced on the day in the form of a set, respectively recorded as ; Extract the importance evaluation coefficient of each keyword in the keyword tag of each data item newly introduced on the day, determine the importance weight of each keyword in the keyword tag of each data item newly introduced on the day, and then map the importance weight of the corresponding set element; By formula Analyze the correlation coefficient between each newly introduced data item and the keyword label of each sub-database in the target database, where They represent the newly introduced The keyword tag set of each data item is the same as the keyword tag set of the target database. The number of elements in the intersection and union of the keyword tag sets of the sub-libraries, Indicates the newly introduced The keyword tag set of each data item is the same as the keyword tag set of the target database. The keyword tag set of the sub-library The importance weights of the intersection elements, is the number of each intersection element, ; According to the predefined subject words corresponding to the replacement vectors stored in the WEB cloud, the subject labels of each sub-database in the target database and the subject labels of each data item newly introduced on the day are represented in the form of vectors, which are recorded as , according to the formula Analyze the correlation coefficient between each newly introduced data item and the subject label of each sub-database in the target database; Said and The specific analysis process includes: recording the time span corresponding to the time label of each sub-database in the target database and the time span corresponding to the time label of each data item newly introduced on the day as , according to the calculation model Analyze the correlation coefficient between each newly introduced data item and the time label of each sub-database in the target database, where ; Similarly, compare the spatial labels corresponding to the regional categories of each sub-database in the target database and the spatial labels corresponding to the regional categories of each data item newly introduced on the day, and analyze the correlation coefficients of the spatial labels of each data item newly introduced on the day and each sub-database in the target database based on the inclusion relationship between the regional categories; The parsing of the text content of each data item and generating its corresponding index tag includes: removing each stop word involved in the text content of each data item according to a predefined stop word list stored in the WEB cloud, sorting the text word set of each data item after word segmentation, retrieving the word frequency of each word in the text word set compared with its data item and the inverse document frequency of the word in the target database, taking the product of the two as the importance evaluation coefficient, screening out each word whose importance evaluation coefficient is greater than or equal to a preset importance evaluation coefficient reaching standard threshold, recording as each keyword, thereby obtaining each keyword in the text word set of each data item; Determine the part-of-speech category of each keyword in each data item text word set, compare the semantic similarity between keywords in the same part-of-speech category in each data item text word set, match semantically similar keyword pairs in the same part-of-speech category, compare the importance evaluation coefficient values ​​between keywords in the semantically similar keyword pairs, eliminate the keywords corresponding to the one with the smaller importance evaluation coefficient value in the semantically similar keyword pairs, and retain the keywords belonging to each part of speech in each data item text word set as keyword labels for the corresponding data items; According to the word list related to each text topic stored in the WEB cloud, the word distribution probability of each data item text word set relative to each text topic is retrieved, and the text topic corresponding to the maximum word distribution probability is selected as the topic label of the data item, thereby obtaining the topic label of each data item; Retrieving the time attribute words and the spatial location attribute words in the text content of each data item, and generating the time label and the spatial label of each data item in the corresponding preset standard format; Integrate keyword tags, topic tags, time tags, and space tags to generate index tags for each data item; The searching of the word frequency of each word in the text word set compared with its data item and the inverse document frequency of each word in the target database includes: counting the occurrence frequency of each word in the text word set in its data item text content, performing a ratio analysis with the total number of words in the data item text word set, and obtaining the word frequency of each word in the text word set compared with its data item; The number of archived data items in the target database whose text content includes a word in the text word set as of the previous day is counted, and a ratio analysis is performed between the number of archived data items and the total number of archived data items in the target database as of the previous day. The natural logarithm of the ratio analysis result is used as the inverse document frequency of the word in the text word set compared with the target database, and then the inverse document frequency of each word in the text word set compared with the target database is obtained.

2. The method for intelligent analysis of massive information based on big data according to claim 1, characterized in that: The screening of each high-frequency data item and each low-frequency data item in the target database includes: calculating the time span between each retrieval date and the last retrieval date of each archived data item in the target database as of a preset time point on the current day, obtaining a reference retrieval period for each archived data item in the target database by average calculation, and analyzing the retrieval frequency of each archived data item in the target database in combination with the number of retrievals for each archived data item in the target database as of the preset time point on the current day. , is the number of each archived data item in the target database, ; The high-frequency retrieval status of data items stored in the WEB cloud corresponds to the retrieval frequency threshold The retrieval frequency threshold corresponding to the low-frequency retrieval state , will satisfy The archived data items of the condition are regarded as high-frequency data items, which will satisfy The archived data items of the conditions are used as low-frequency data items to filter out the high-frequency data items and low-frequency data items in the target database.

3. The method for intelligent analysis of massive information based on big data according to claim 1, characterized in that: The index tag optimization processing for each high-frequency data item in the target database includes: extracting the input search content of each search of each high-frequency data item in the target database as of a preset time point on the same day, finding the newly added words and semantic deviation degree of the input search content of each search of the high-frequency data item relative to the index tag, analyzing the optimization demand coefficient of the index tag of each high-frequency data item in the target database, identifying that the index tag of each high-frequency data item in the target database whose optimization demand coefficient is greater than or equal to a preset optimization demand coefficient warning threshold has optimization demand, and selecting a preset number of valid words from the newly added words of the index tag of the input search content of each search, and adding them to the index tag.

4. The method for intelligent analysis of massive information based on big data according to claim 1, characterized in that: The storage location optimization processing for each high-frequency data item in the target database includes: extracting the index tags of each high-frequency data item in the target database after the optimization processing, analyzing the comprehensive correlation index between each high-frequency data item in the target database and each sub-database, and then screening each associated sub-database corresponding to the optimized index tags of each high-frequency data item in the target database, comparing with each associated sub-database stored corresponding to each high-frequency data item, and determining each high-frequency data item corresponds to each newly added associated sub-database; Extract the corresponding call associated sub-databases of each high-frequency data item in the target database as of the preset time point on the day, evaluate the corresponding search activity of each associated sub-database in which each high-frequency data item in the target database has been stored, and take the associated sub-databases whose search activity is greater than or equal to the preset search activity warning threshold as the retained associated sub-databases; The newly added associated sub-libraries and the retained associated sub-libraries corresponding to the high-frequency data items are integrated to obtain the associated sub-libraries stored corresponding to the high-frequency data items after optimization processing.

5. The method for intelligent analysis of massive information based on big data according to claim 1, characterized in that: The data cleaning process for the low-frequency data database includes: determining the number of oversaturated data items in the low-frequency data database, extracting the retrieval frequency of each low-frequency data item in the low-frequency data database, analyzing the retrieval value coefficient of each low-frequency data item in the low-frequency data database in combination with the import date of each low-frequency data item into the target database, and cleaning each low-frequency data item whose retrieval value coefficient is ranked after the number of oversaturated data items.

Images