Public Opinion Security Analysis Method, Device and Electronic Device Based on Event Relevance

Abstract

The present invention discloses a public opinion security analysis method, device and electronic device based on event correlation. The method includes: classifying public opinion data into corresponding public opinion classification sets; performing dimensionality reduction processing on the public opinion data in the public opinion classification sets, and based on a first time window, searching for the historical event classification set to which the processed public opinion data belongs, and classifying the public opinion data into the historical event classification set to which it belongs; performing correlation matching on the public opinion data in the same historical event classification set, and classifying the matched public opinion data into the negative public opinion set; performing public opinion security analysis based on the public opinion classification set, the historical event classification set and the negative public opinion set. The present invention can classify, trace the origin, ferment and re-ferment public opinion; thus quickly restoring the complete occurrence and fermentation process of public opinion events, preventing potential public opinion risks, ensuring public opinion security, and maintaining the image of enterprises and society.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, electronic device and computer-readable medium for public opinion security analysis based on event correlation. Background Art

[0002] With the rapid development and popularization of the Internet, people have become accustomed to expressing their opinions or comments on social hot spots and public affairs through the Internet; at the same time, various forms of self-media and social platforms have also emerged, such as public accounts, Weibo, etc. When social events and social problems occur, the public often quickly use media platforms to understand the causes and development of the events, and then express their opinions through online media. These opinions have an impact on the development of the events that cannot be ignored, thus generating public opinion. In the process of the emergence and spread of public opinion, there are often some high-repetition accounts, content forwarding accounts, etc. that edit and distort the events, which seriously affect the public image of enterprises, individuals, and even society. Therefore, public opinion security is crucial to enterprises and society.

[0003] At present, due to the rapidity, extensiveness and strong interactivity of network communication, online public opinion often grows explosively and its forms are complex and diverse, resulting in large amounts of public opinion information data, numerous contents, difficulty in tracing the source, diverse identities of public opinion publishers, flexible forwarding, and many points of concern, making it impossible to conduct a rapid and effective analysis of public opinion. Summary of the invention

[0004] In view of this, the main purpose of the present invention is to propose a public opinion security analysis method, device, electronic device and computer-readable medium based on event correlation, in order to at least partially solve at least one of the above-mentioned technical problems.

[0005] In order to solve the above technical problems, the first aspect of the present invention proposes a public opinion security analysis method based on event correlation, the method comprising:

[0006] Classify the public opinion data into corresponding public opinion classification sets;

[0007] Perform dimensionality reduction processing on the public opinion data in the public opinion classification set, and search the historical event classification set to which the processed public opinion data belongs based on the first time window, and classify the public opinion data into the historical event classification set to which it belongs;

[0008] Perform correlation matching on the public opinion data in the same historical event classification set, and classify the matching public opinion data into the negative public opinion set;

[0009] Public opinion security analysis is performed based on the public opinion classification set, the historical event classification set and the negative public opinion set.

[0010] According to a preferred embodiment of the present invention, the dimensionality reduction processing of the public opinion data in the public opinion classification set and the search for the historical event classification set to which the processed public opinion data belongs based on the first time window include:

[0011] Perform binary conversion on the public opinion data in the public opinion classification set, and split the converted string to obtain public opinion segments;

[0012] Search for the strings to be compared related to each public opinion segment within the first time window;

[0013] Compare the string where the public opinion segment is located with the string to be compared bit by bit to determine the similar string of the public opinion data;

[0014] Determine the historical event classification set to which the public opinion data belongs according to the similar string of the public opinion data.

[0015] According to a preferred embodiment of the present invention, the method further includes:

[0016] Perform dimensionality reduction processing on the newly added public opinion data, search for the historical event classification set to which the processed newly added public opinion data belongs based on the first time window, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event fermentation alarm message;

[0017] If the historical event classification set to which the processed newly added public opinion data belongs cannot be found within the first time window, replace the first time window with the second time window, search for the historical event classification set to which the processed newly added public opinion data belongs based on the second time window, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event secondary fermentation alarm message;

[0018] Wherein: the second time window is earlier than the first time window.

[0019] According to a preferred embodiment of the present invention, the comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the first time window is different from the comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the second time window.

[0020] According to a preferred embodiment of the present invention, the method further includes: searching for and displaying the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set according to the user search information.

[0021] According to a preferred embodiment of the present invention, the searching for and displaying the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set according to the user search information includes:

[0022] Match the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set based on multiple dimensions according to the user search information;

[0023] Determine the matching degree of each target public opinion according to the weight value of each dimension;

[0024] Display each target public opinion according to the matching degree of the target public opinion.

[0025] According to a preferred embodiment of the present invention, before displaying each target public opinion according to the matching degree of the target public opinion, the method further includes:

[0026] Adjust the matching degree of the target public opinion according to the historical target public opinion matching the user's search information; or:

[0027] Adjust the matching degree of the target public opinion according to the final public opinion selected by the user from the historical target public opinion.

[0028] According to a preferred embodiment of the present invention, the matching of the target public opinion from the public opinion classification set, the historical event classification set, and the negative public opinion set based on multiple dimensions according to the user's search information includes:

[0029] Match the user's search information based on multiple dimensions from the public opinion classification set, the historical event classification set, and the negative public opinion set respectively to obtain classified target public opinion, event target public opinion, and negative target public opinion.

[0030] According to a preferred embodiment of the present invention, the method further includes:

[0031] Construct a user profile according to the user's historical search;

[0032] Push information to the user based on the user profile.

[0033] According to a preferred embodiment of the present invention, the classification processing of the public opinion data into the corresponding public opinion classification set includes:

[0034] Conduct a correlation analysis on the public opinion data to obtain relevant public opinion data;

[0035] Classify the relevant public opinion data according to the text attributes thereof;

[0036] Judge the emotional type of the classified public opinion data based on the positive and negative sentiment corpus;

[0037] Process the sentiment-classified public opinion data using a classification model to obtain the public opinion classification set where the public opinion data is located.

[0038] According to a preferred embodiment of the present invention, if the emotional type of the classified public opinion data is negative, issue a negative public opinion alert.

[0039] According to a preferred embodiment of the present invention, the public opinion data includes: media public opinion data and social network public opinion data, and the method further includes:

[0040] Perform correlation matching on the social network public opinion data, and classify the matched social network public opinion data into the negative public opinion set.

[0041] To solve the above technical problems, a second aspect of the present invention provides an opinion security analysis device based on event correlation, and the device includes:

[0042] A classification module for classifying public opinion data into corresponding public opinion classification sets;

[0043] A dimensionality reduction search module for performing dimensionality reduction processing on the public opinion data in the public opinion classification set, and searching for the historical event classification set to which the processed public opinion data belongs based on the first time window, and classifying the public opinion data into the historical event classification set to which it belongs;

[0044] A first matching module for performing correlation matching on the public opinion data in the same historical event classification set, and classifying the matched public opinion data into the negative public opinion set;

[0045] An analysis module for performing opinion security analysis based on the public opinion classification set, the historical event classification set, and the negative public opinion set.

[0046] According to a preferred embodiment of the present invention, the dimensionality reduction search module includes:

[0047] A segmentation module for performing binary conversion on the public opinion data in the public opinion classification set, and segmenting the converted string to obtain public opinion segments;

[0048] A sub-search module for searching for strings to be compared related to each public opinion segment within the first time window;

[0049] A comparison module for comparing the string where the public opinion segment is located with the string to be compared bit by bit to determine the similar string of the public opinion data;

[0050] A sub-determination module for determining the historical event classification set to which the public opinion data belongs according to the similar string of the public opinion data.

[0051] According to a preferred embodiment of the present invention, the device further includes:

[0052] A first dimensionality reduction search module for performing dimensionality reduction processing on new public opinion data, and searching for the historical event classification set to which the processed new public opinion data belongs based on the first time window, classifying the new public opinion data into the historical event classification set to which it belongs, and sending out an event fermentation alarm message;

[0053] The second dimensionality reduction search module is configured to, if the historical event classification set to which the processed new public opinion data belongs is not found within the first time window, replace the first time window with the second time window, search for the historical event classification set to which the processed new public opinion data belongs based on the second time window, classify the new public opinion data into the historical event classification set to which it belongs, and send an event secondary fermentation alarm message;

[0054] Wherein: the second time window is earlier than the first time window.

[0055] According to a preferred embodiment of the present invention, the comparison mechanisms of the first dimensionality reduction search module and the second dimensionality reduction search module are different.

[0056] According to a preferred embodiment of the present invention, the device further includes: a search and display module, configured to search for and display target public opinions from the public opinion classification set, the historical event classification set, and the negative public opinion set according to the user's search information.

[0057] According to a preferred embodiment of the present invention, the search and display module includes:

[0058] A multi-dimensional matching module, configured to match target public opinions from the public opinion classification set, the historical event classification set, and the negative public opinion set based on multiple dimensions according to the user's search information;

[0059] A determination module, configured to determine the matching degree of each target public opinion according to the weight value of each dimension;

[0060] A display module, configured to display each target public opinion according to the matching degree of the target public opinion.

[0061] According to a preferred embodiment of the present invention, the device further includes:

[0062] An adjustment module, configured to adjust the matching degree of the target public opinion according to the historical target public opinion matching the user's search information; or: adjust the matching degree of the target public opinion according to the final public opinion selected by the user from the historical target public opinions.

[0063] According to a preferred embodiment of the present invention, the multi-dimensional matching module includes:

[0064] A sub-matching module, configured to match the user's search information based on multiple dimensions from the public opinion classification set, the historical event classification set, and the negative public opinion set respectively, to obtain classified target public opinions, event target public opinions, and negative target public opinions.

[0065] According to a preferred embodiment of the present invention, the device further includes:

[0066] A construction module, configured to construct a user portrait according to the user's historical search;

[0067] A push module for pushing information to users based on user portraits.

[0068] According to a preferred embodiment of the present invention, the classification module includes:

[0069] A correlation analysis module for performing a correlation analysis on public opinion data to obtain relevant public opinion data;

[0070] A sub-classification module for classifying the relevant public opinion data according to the text attributes thereof;

[0071] A judgment module for judging the sentiment type of the classified public opinion data based on a positive and negative sentiment corpus;

[0072] A model processing module for processing the sentiment-classified public opinion data by using a classification model to obtain a public opinion classification set where the public opinion data is located.

[0073] According to a preferred embodiment of the present invention, the device further includes: an alarm module for issuing a negative public opinion alarm if the sentiment type of the classified public opinion data is negative.

[0074] According to a preferred embodiment of the present invention, the public opinion data includes: media public opinion data and social network public opinion data, and the device further includes:

[0075] A second matching module for performing a correlation match on the social network public opinion data and classifying the matched social network public opinion data into a negative public opinion set.

[0076] To solve the above technical problems, a third aspect of the present invention provides an electronic device, including:

[0077] A processor; and

[0078] A memory storing computer-executable instructions, and when the computer-executable instructions are executed, the processor executes the above method.

[0079] To solve the above technical problems, a fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the above method is implemented.

[0080] The present invention classifies public opinion data into corresponding public opinion classification sets, performs dimensionality reduction processing on the public opinion data in the public opinion classification sets, and based on the first time window, searches for the historical event classification set to which the processed public opinion data belongs, and classifies the public opinion data into the historical event classification set to which it belongs. For the public opinion data in the same historical event classification set, perform correlation matching, and classify the matched public opinion data into the negative public opinion set; according to the public opinion classification set, the public opinion category can be analyzed, according to the historical event classification set, the origin, fermentation and secondary fermentation of the public opinion can be analyzed, and according to the negative public opinion set, the repeated accounts and forwarding accounts during the generation process of the public opinion can be tracked and analyzed; thus, the complete occurrence and fermentation process of the public opinion event can be quickly restored. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] In order to make the technical problems solved by the present invention, the technical means adopted and the technical effects obtained more clear, the specific embodiments of the present invention will be described in detail below with reference to the drawings. However, it should be noted that the drawings described below are only the drawings of the exemplary embodiments of the present invention, and those skilled in the art can obtain the drawings of other embodiments without creative work based on these drawings.

[0082] Figure 1 is a schematic flowchart of a public opinion security analysis method based on event correlation according to an embodiment of the present invention;

[0083] Figure 2 is a schematic flowchart of dimensionality reduction processing on the public opinion data in the public opinion classification set according to an embodiment of the present invention, and searching for the historical event classification set to which the processed public opinion data belongs based on the first time window;

[0084] Figure 3 is a schematic flowchart of searching and displaying target public opinion from the public opinion classification set, historical event classification set and negative public opinion set according to user search information according to an embodiment of the present invention;

[0085] Figure 4 is a schematic structural framework diagram of a public opinion security analysis device based on event correlation according to an embodiment of the present invention;

[0086] Figure 5 is a schematic block diagram of the structure of an exemplary embodiment of an electronic device according to the present invention;

[0087] Figure 6 is a schematic diagram of an embodiment of a computer-readable medium according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0088] Exemplary embodiments of the present invention will now be described more fully with reference to the accompanying drawings. Although the exemplary embodiments can be implemented in many specific ways, the present invention should not be construed as being limited to the embodiments set forth herein. On the contrary, these exemplary embodiments are provided to make the content of the present invention more complete and to more conveniently convey the inventive concept to those skilled in the art.

[0089] On the premise of conforming to the technical concept of the present invention, the structures, performances, effects or other features described in a specific embodiment can be combined with one or more other embodiments in any suitable manner.

[0090] In the process of introducing specific embodiments, the detailed descriptions of the structures, performances, effects or other features are for those skilled in the art to fully understand the embodiments. However, it does not exclude that those skilled in the art can implement the present invention with technical solutions that do not contain the above-mentioned structures, performances, effects or other features under specific circumstances.

[0091] The flowcharts in the accompanying drawings are only exemplary flow demonstrations, and do not mean that all the contents, operations and steps in the flowcharts must be included in the solutions of the present invention, nor does it mean that they must be executed in the order shown in the figures. For example, some operations / steps in the flowchart can be decomposed, some operations / steps can be combined or partially combined, etc. Without departing from the gist of the present invention, the execution order shown in the flowchart can be changed according to the actual situation.

[0092] The boxes in the accompanying drawings Figure 1 generally represent functional entities, and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0093] The same reference numerals in the drawings denote the same or similar elements, components or parts. Therefore, the repeated descriptions of the same or similar elements, components or parts may be omitted hereinafter. It should also be understood that although the first, second, third, etc. attributives indicating numbers may be used herein to describe various devices, elements, components or parts, these devices, elements, components or parts should not be limited by these attributives. That is, these attributives are only used to distinguish one from another. For example, the first device can also be called the second device without departing from the essential technical solution of the present invention. In addition, the terms "and / or", "or / and" mean all combinations including any one or more of the listed items.

[0094] Please refer to Figure 1 , Figure 1A public opinion security analysis method based on event correlation provided by the present invention is as follows Figure 1 As shown, the method includes:

[0095] S1. Classify the public opinion data and assign it to the corresponding public opinion classification set;

[0096] Among them, the public opinion data may include: media public opinion data from official media, social network public opinion data from online social platforms, media public opinion data from self-media, and so on.

[0097] In this embodiment, the public opinion data is processed in text form. Therefore, before this step, public opinion texts can be obtained from official media, self-media, and online social platforms first. Then, this step classifies the public opinion texts according to attributes such as the length, time, sentiment attribute, term frequency-inverse document frequency, and source of the text. Exemplarily, this step may include:

[0098] S11. Conduct a correlation analysis on the public opinion data to obtain relevant public opinion data;

[0099] For example: Similar or relevant texts in the public opinion text can be identified based on a semantic model, and the similar or relevant public opinion texts are used as relevant public opinion data.

[0100] S12. Classify the relevant public opinion data according to its text attributes;

[0101] Exemplarily, the text attributes may include: attributes such as the length, time, term frequency-inverse document frequency, and source of the text. Taking the processing according to the term frequency-inverse document frequency as an example, first input the public opinion text into the Analyzer tokenizer for tokenization, process the public opinion text into tokens, filter out the stop words in the tokens to obtain valid tokens, then perform term frequency-inverse document frequency processing on the public opinion text based on the valid tokens to obtain the keywords of each public opinion text, and classify the public opinion text based on the keywords. The step of filtering out the stop words in the tokens to obtain valid tokens may include: judging whether there are stop words in the tokens. If there are, filter out the stop words from the tokens to obtain valid tokens. If not, take all the tokens as valid words.

[0102] S13. Judge the sentiment type of the classified public opinion data based on the positive and negative sentiment corpus;

[0103] In this embodiment, the positive and negative sentiment corpus is a corpus containing positive and negative text keywords constructed based on historical corpora. Through this corpus, word segmentation and sentiment analysis can be performed on new situations in real time. In this step, the word segmentation after filtering stop words is compared with the positive and negative text keywords in the positive and negative sentiment corpus, the number of positive and negative words is counted, and the sentiment type of the public opinion text is determined according to the ratio of positive and negative words. For example: if the ratio of positive and negative words is greater than the first threshold, the sentiment type of the public opinion text is positive; if the ratio of positive and negative words is less than the second threshold, the sentiment type of the public opinion text is negative.

[0104] Furthermore, if this step determines that the sentiment type of the public opinion text is negative, a negative public opinion alert is issued to facilitate the timely monitoring, attention, and handling of public opinion, and to control risk public opinion at an early stage.

[0105] S14. Use a classification model to process the sentiment-classified public opinion data to obtain the public opinion classification set where the public opinion data is located.

[0106] Exemplarily, public opinion texts with the same initial classification and sentiment classification, and / or the word segmentation of the public opinion text corresponding thereto are input into a pre-trained classification model, and the public opinion is finally classified according to the output result of the model to obtain the public opinion classification set where each public opinion text and / or word segmentation of the public opinion is located. Among them: one public opinion classification set corresponds to a public opinion category of a certain sentiment type, and the sentiment categories may include: support, praise, opposition, criticism, etc. The public opinion categories may include: technology, military, entertainment, politics, news, etc.

[0107] Furthermore, the same label can be assigned to the public opinion texts and / or word segmentations of the same category for distinction. A public opinion classification database is constructed through the labeled public opinion texts and / or word segmentations of the public opinion to facilitate later use. To facilitate subsequent searching based on the public opinion classification database, an inverted index can be established for the public opinion data while constructing the public opinion classification database.

[0108] In addition, since high-repetition accounts, content forwarding accounts, etc. mainly come from self-media and social network platforms, therefore, in this step, the self-media public opinion data and social network public opinion data can also be matched for relevance, and the matched social network public opinion data is classified into the negative public opinion set for key attention. Exemplarily, the relevance can be matched through keywords and / or public opinion release accounts.

[0109] S2. Perform dimensionality reduction processing on the public opinion data in the public opinion classification set, and based on the first time window, find the historical event classification set to which the processed public opinion data belongs, and classify the public opinion data into the historical event classification set to which it belongs;

[0110] Considering that a large amount of public opinion data will affect the processing efficiency, in this embodiment, the public opinion data is dimensionally reduced, and relevant historical events are searched based on a sliding event window to realize the traceability analysis of public opinion. Among them: the dimensionality reduction processing may include: PCA, ICA, LDA, ISOMAP, LLE, etc. For the convenience of processing in this embodiment, the dimensionality reduction processing converts the public opinion data in the public opinion classification concentration into a binary string, as follows Figure 2 shown: This step may include:

[0111] S21. Perform binary conversion on the public opinion data in the public opinion classification concentration, and split the converted string to obtain public opinion segments;

[0112] Exemplarily, if the public opinion classification concentration only includes public opinion texts, the public opinion texts are first input into the Analyzer tokenizer for tokenization processing, the public opinion texts are processed into tokens, the stop words in the tokens are filtered, and the valid tokens are obtained. The valid tokens are converted into 64-bit binary strings. If the public opinion classification concentration includes both public opinion texts and the corresponding valid tokens of the public opinion texts, the valid tokens are directly converted into 64-bit binary strings.

[0113] In this embodiment, in order to further improve the processing speed, the string obtained by dimensionality reduction can be split, and one of the public opinion segments is used to search for relevant historical event classification sets, and then all the public opinion data of the relevant historical event classification sets are recalled and compared with the entire string, thereby greatly reducing the calculation amount of the comparison and improving the processing speed.

[0114] Among them: the string can be split in an equal-length splitting manner. For example, a 64-bit binary string is equally split into 4 16-bit strings, and each equal split string is a public opinion segment. Optionally, each public opinion segment can be cached in the redis database, and the public opinion segments obtained by dimensionality reduction and splitting of the same public opinion text are marked with the same identifier for subsequent use.

[0115] S22. Search for strings to be compared related to each public opinion segment within the first time window;

[0116] The time window can be set according to actual needs. In this embodiment, a sliding time window can be adopted. Therefore, the first and second are used to distinguish different time windows. For example, the first time window can be set within two days from now, and the second time window can be set within two to four days from now.

[0117] Exemplarily, in this step, each sentiment segment after dimensionality reduction and segmentation of the same sentiment text can be obtained first, and the strings to be compared related to the sentiment segment that occurred within the first time window are searched according to each sentiment segment. Among them: Searching for the strings to be compared that occurred within the first time window according to each sentiment segment may include: obtaining the string after dimensionality reduction processing of the sentiment text that occurred within the first time window, denoted as the string to be matched, and performing correlation matching between each sentiment segment and each string to be matched to obtain the strings to be compared. Taking the example that the sentiment text is dimensionally reduced and segmented into 4 sentiment segments of 16-bit strings, first obtain 4 16-bit strings after dimensionality reduction and segmentation of the same sentiment text from the redis database according to the identifier, and obtain the strings to be matched after dimensionality reduction processing of the sentiment text that occurred within the first time window from the redis database. For example, 10 64-bit strings to be matched are obtained, and each 16-bit string is correlated with each 64-bit string to be matched, and the successfully matched strings to be matched are used as the strings to be compared.

[0118] S23. Compare the string where the sentiment segment is located with the strings to be compared bit by bit to determine the strings with similar sentiment data.

[0119] Among them: The string where the sentiment segment is located is the string before segmentation. In this example, the string where the sentiment segment is located is a 64-bit string obtained by binary conversion of the sentiment text after word segmentation. Then, in this step, this 64-bit string is compared with each 64-bit string to be compared bit by bit. If the number of different bits is less than a predetermined value (such as 4), this 64-bit string to be compared is used as the string with similar sentiment data. Exemplarily, the simhash method and the sentiment analysis result can be used for comparison during the comparison process.

[0120] S24. Determine the historical event classification set to which the sentiment data belongs according to the strings with similar sentiment data.

[0121] Exemplarily, the historical event classification sets to which each string with similar sentiment data belongs can be searched first, and the historical event classification set with the most strings with similar sentiment data is used as the historical event classification set to which the sentiment data belongs.

[0122] Subsequently, the sentiment data (such as: sentiment text, the string corresponding to the sentiment text, the word segmentation corresponding to the sentiment text) is classified into the historical event classification set to which it belongs. If there is no string with similar sentiment data, a new historical event classification set is created for this sentiment data.

[0123] Furthermore, the dimensionality reduction processing of the newly added sentiment data can be performed in real time, and relevant historical events can be searched based on the sliding event window to realize the early warning of the fermentation and secondary fermentation of the sentiment. Therefore, the method may further include:

[0124] S201. Perform dimensionality reduction processing on the newly added public opinion data, search for the historical event classification set to which the processed newly added public opinion data belongs based on the first time window, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event fermentation alarm message.

[0125] S202. If the historical event classification set to which the processed newly added public opinion data belongs cannot be found within the first time window, replace the preset time window with the second time window, search for the historical event classification set to which the processed newly added public opinion data belongs based on the second time window, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event secondary fermentation alarm message.

[0126] Wherein: the second time window is earlier than the first time window. The search method for the historical event classification set to which the newly added public opinion data belongs can refer to steps S21 - S24.

[0127] Furthermore, in order to make the event traceability analysis clearer and distinguish between the first fermentation and the second fermentation, the comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the first time window is different from the comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the second time window. Among them: the comparison mechanism can include: comparison methods, such as: the simhash method, sentiment analysis results, etc., for example, content: such as: text, string, account, keyword, and so on.

[0128] In addition, the same label can be assigned to the public opinion text, public opinion word segmentation, and string in the same historical event classification set for distinction. A historical event classification database is constructed through the labeled public opinion text, public opinion word segmentation, and string, which is convenient for later use. In order to facilitate subsequent search based on the historical event classification database, an inverted index can be established for the public opinion data while constructing the historical event classification database.

[0129] S3. Perform correlation matching on the public opinion data in the same historical event classification set, and classify the matched public opinion data into the negative public opinion set.

[0130] In this embodiment, correlation matching can be performed through at least one of a string, a keyword, and a public opinion release account. By establishing a negative public opinion set, the dimension of public opinion security analysis based on event correlation can be supplemented.

[0131] S4. Perform public opinion security analysis based on the public opinion classification set, historical event classification set, and the negative public opinion set.

[0132] Exemplarily, the sentiment category can be analyzed according to the sentiment classification set. The origin, fermentation, and secondary fermentation of the sentiment can be analyzed according to the historical event classification set. The repeated accounts, forwarding accounts, etc. during the generation process of the sentiment can be tracked and analyzed according to the negative sentiment set, so as to quickly restore the complete occurrence and fermentation process of the sentiment event.

[0133] Further, the embodiments of the present invention can also search and display the user input information based on the above-mentioned sentiment classification set, historical event classification set, and negative sentiment set to ensure the security and reliability of the information displayed to the user. Based on this, the method further includes:

[0134] S5. Search and display the target sentiment from the sentiment classification set, historical event classification set, and negative sentiment set according to the user search information.

[0135] Wherein: the user search information can be information such as keywords, texts input by the user. Exemplarily, as Figure 3 shown, this step may include:

[0136] S51. Match the target sentiment from the sentiment classification set, historical event classification set, and negative sentiment set based on multiple dimensions according to the user search information;

[0137] Wherein: the dimensions are used to match the target sentiment from different perspectives. Exemplarily, the target sentiment can be matched from dimensions such as the matching degree of the article title, content matching degree, popularity, tf-idf value, originality, author portrait, etc. Exemplarily, multi-dimensional matching can be performed through matching rules or matching models.

[0138] S52. Determine the matching degree of each target sentiment according to the weight value of each dimension;

[0139] In this embodiment, the weight of each dimension can be set in advance, and the matching degree of each target sentiment is calculated based on the score of each dimension of the target sentiment and the weight of the corresponding dimension.

[0140] In a preferred example, the matching degree of the target public opinion can also be adjusted by integrating the user's historical search results, so that the display result is closer to the user's needs and the user experience is improved. Then, in this step, the matching degree of the target public opinion can also be adjusted according to the historical target public opinion that matches the user's search information; specifically, obtain the historical target public opinion that matches the user's search information; compare the historical target public opinion with the target public opinion; and adjust the matching degree (i.e., the display order of the target public opinion) of the target public opinion according to the comparison result. Exemplarily, the matching degree of the target public opinion is adjusted according to the number of the same historical target public opinions. For example, take the number of the same historical target public opinions as a dimension, set the corresponding weight, and adjust the matching degree based on this. Or directly adjust the matching degree proportionally according to the number of the same historical target public opinions.

[0141] In addition, the matching degree of the target public opinion can also be adjusted according to the final public opinion selected by the user from the displayed historical target public opinions. Then, in this step, it can also: obtain the final public opinion selected by the user from the historical target public opinions; compare the final public opinion with the target public opinion; and adjust the matching degree of the target public opinion according to the comparison result.

[0142] S53. Display each target public opinion according to the matching degree of the target public opinion.

[0143] For example: Display the target public opinions from top to bottom according to the matching degree in the form of a drop-down list. Or set different display areas, and display the target public opinions according to the significant level of the area based on the matching degree. For example: Display the target public opinion with the largest matching degree in the display area with the first significant level, and display the target public opinion with the second largest matching degree in the display area with the second significant level. Among them: The display level reflects the speed at which the user discovers the content in the display area. For example: The content in the middle area of the display screen is most easily discovered by the user.

[0144] Furthermore, the embodiments of the present invention can also analyze the user attributes based on the user's historical searches in the above-mentioned public opinion classification set, historical event classification set, and negative public opinion set, construct a user portrait, and push information to the user. Based on this, the method further includes:

[0145] S6. Construct a user portrait according to the user's historical searches, and push information to the user based on the user portrait.

[0146] Among them: The user's historical searches may include information such as keywords and texts input by the user, and may also include the texts finally selected by the user in the search results.

[0147] Compared with the prior art, the present invention has at least the following beneficial effects:

[0148] 1. It can classify, trace the source, ferment, and re-ferment a large amount of public opinion and security intelligence in the system database, restore the complete occurrence and fermentation process of events, facilitate the staff to prevent the events that are about to ferment in a timely manner, and ensure the public opinion security of enterprises and society.

[0149] 2. Perform dimensionality reduction processing on public opinion data, segment the dimensionality-reduced public opinion data, use one of the public opinion segments to search for relevant strings to be compared, and then recall the entire string where the public opinion segment is located and compare it with each string to be compared, thereby greatly reducing the computational complexity of the comparison and improving the processing speed.

[0150] 3. In the analysis process, the large number of algorithm recalls and large computational complexity lead to a long time consumption. Since the computational complexity cannot be reduced, the core idea is to reduce the recall time, trading space for time. The recall of each 16-bit substring can be directly obtained from Redis without querying, which greatly reduces the comparison time.

[0151] 4. Build a structured large-scale intelligence library, classify the text according to attributes such as text length, time, sentiment attribute, term frequency-inverse document frequency, source, etc., and build an inverted index while storing it in the library.

[0152] 5. Calculate weighted scores for information such as the matching degree of article titles, content matching degree, popularity, tf-idf value, originality, and author portrait. At the same time, analyze the user's historical search habits, find high-frequency keywords, conduct a secondary search on the first search results and calculate scores. Calculate the scores of the two searches to obtain the final article presentation order.

[0153] 6. Search and display the user input information based on the above-mentioned public opinion classification set, historical event classification set, and negative public opinion set to ensure the security and reliability of the information presented to the user.

[0154] 7. Based on different comparison mechanisms for new public opinion in different time windows, it can distinguish the first, second, and multiple repeated fermentations of events, and accurately trace the propagation process and link of each event.

[0155] 8. It can analyze and judge behaviors such as one person having multiple accounts and forwarding during the time propagation process, find the key communication population, and at the same time reduce the error of public opinion analysis.

[0156] 9. By extracting the user's search history, build a user portrait, realize personalized user recommendations for each user, and optimize the user search experience.

[0157] 10. It can deduplicate a large amount of public opinion and security texts; it can judge whether the text is positive or negative according to the attributes of public opinion and security texts.

[0158] Figure 4This is an opinion sentiment security analysis device based on event correlation according to the present invention. As Figure 4 shown, the device includes:

[0159] A classification module 41 for classifying public opinion data into corresponding public opinion classification sets;

[0160] A dimensionality reduction search module 42 for performing dimensionality reduction processing on the public opinion data in the public opinion classification set, searching for the historical event classification set to which the processed public opinion data belongs based on the first time window, and classifying the public opinion data into the historical event classification set to which it belongs;

[0161] A first matching module 43 for performing correlation matching on the public opinion data in the same historical event classification set, and classifying the matched public opinion data into the negative public opinion set;

[0162] An analysis module 44 for performing public opinion sentiment security analysis based on the public opinion classification set, the historical event classification set, and the negative public opinion set.

[0163] In an implementation manner, the dimensionality reduction search module 42 includes:

[0164] A segmentation module for performing binary conversion on the public opinion data in the public opinion classification set, and segmenting the converted string to obtain public opinion segments;

[0165] A sub-search module for searching for strings to be compared related to each public opinion segment within the first time window;

[0166] A comparison module for performing bit-by-bit comparison between the string where the public opinion segment is located and the string to be compared to determine the similar string of the public opinion data;

[0167] A sub-determination module for determining the historical event classification set to which the public opinion data belongs according to the similar string of the public opinion data.

[0168] Furthermore, the device further includes:

[0169] A first dimensionality reduction search module for performing dimensionality reduction processing on the newly added public opinion data, searching for the historical event classification set to which the processed newly added public opinion data belongs based on the first time window, classifying the newly added public opinion data into the historical event classification set to which it belongs, and sending out an event fermentation alarm message;

[0170] A second dimensionality reduction search module for, if the historical event classification set to which the processed newly added public opinion data belongs cannot be found within the first time window, replacing the first time window with a second time window, searching for the historical event classification set to which the processed newly added public opinion data belongs based on the second time window, classifying the newly added public opinion data into the historical event classification set to which it belongs, and sending out an event secondary fermentation alarm message;

[0171] Wherein, the second time window is earlier than the first time window.

[0172] The comparison mechanisms of the first dimensionality reduction search module and the second dimensionality reduction search module are different.

[0173] Furthermore, the device further includes: a search and display module, configured to search for and display target public opinions from the public opinion classification set, historical event classification set, and negative public opinion set according to user search information.

[0174] In one implementation, the search and display module includes:

[0175] A multi-dimensional matching module, configured to match target public opinions from the public opinion classification set, historical event classification set, and negative public opinion set based on multiple dimensions according to user search information;

[0176] A determination module, configured to determine the matching degrees of the respective target public opinions according to the weight values of each dimension;

[0177] A display module, configured to display the respective target public opinions according to the matching degrees of the target public opinions.

[0178] Furthermore, the device further includes:

[0179] An adjustment module, configured to adjust the matching degrees of the target public opinions according to historical target public opinions matching the user search information; or: adjust the matching degrees of the target public opinions according to the final public opinions selected by the user from the historical target public opinions.

[0180] In one implementation, the multi-dimensional matching module includes:

[0181] A sub-matching module, configured to match the user search information based on multiple dimensions from the public opinion classification set, historical event classification set, and negative public opinion set respectively, to obtain classified target public opinions, event target public opinions, and negative target public opinions.

[0182] Furthermore, the device further includes:

[0183] A construction module, configured to construct a user portrait according to the user's historical searches;

[0184] A push module, configured to push information to the user based on the user portrait.

[0185] In one implementation, the classification module includes:

[0186] A correlation analysis module, configured to perform a correlation analysis on public opinion data to obtain relevant public opinion data;

[0187] A sub-classification module, configured to classify the relevant public opinion data according to the text attributes thereof;

[0188] A judgment module, configured to judge the sentiment type of the classified public opinion data based on a positive and negative sentiment corpus;

[0189] A model processing module, configured to process the classified public opinion data by using a classification model to obtain a public opinion classification set where the public opinion data is located.

[0190] Further, the device further includes: an alarm module, configured to issue a negative public opinion alarm if the sentiment type of the classified public opinion data is negative.

[0191] Further, the public opinion data includes: media public opinion data and social network public opinion data, and the device further includes:

[0192] A second matching module, configured to perform correlation matching on the social network public opinion data, and classify the matched social network public opinion data into a negative public opinion set.

[0193] Those skilled in the art can understand that the various modules in the above device embodiments can be distributed in the device according to the description, or can be correspondingly changed and distributed in one or more devices different from the above embodiments. The modules in the above embodiments can be combined into one module, or further split into multiple sub-modules.

[0194] Next, an embodiment of an electronic device according to the present invention is described. This electronic device can be regarded as an implementation form of the entity for the above method and device embodiments of the present invention. For the details described in the embodiment of the electronic device of the present invention, they should be regarded as a supplement to the above method or device embodiments; for the details not disclosed in the embodiment of the electronic device of the present invention, they can be implemented with reference to the above method or device embodiments.

[0195] Figure 5 It is a structural block diagram of an exemplary embodiment of an electronic device according to the present invention. Figure 5 The shown electronic device is only an example, and should not bring any limitation to the functions and usage scopes of the embodiments of the present invention.

[0196] As Figure 5 shown, the electronic device 500 of this exemplary embodiment is presented in the form of a general-purpose data processing device. The components of the electronic device 500 may include, but are not limited to: at least one processing unit 510, at least one storage unit 520, a bus 530 connecting different electronic device components (including the storage unit 520 and the processing unit 510), a display unit 540, etc.

[0197] Among them, the storage unit 520 stores a computer-readable program, which can be the source program or the code of the read-only program. The program can be executed by the processing unit 510, so that the processing unit 510 executes the steps of various embodiments of the present invention. For example, the processing unit 510 can execute as Figure 1 shown in the steps.

[0198] The storage unit 520 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 5201 and / or a cache storage unit 5202, and may further include a read-only storage unit (ROM) 5203. The storage unit 520 may also include a program / utilities 5204 having a set (at least one) of program modules 5205. Such program modules 4205 include, but are not limited to: operating the electronic device, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0199] The bus 530 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any bus structure in a variety of bus structures.

[0200] The electronic device 500 can also communicate with one or more external devices 100 (such as a keyboard, a display, a network device, a Bluetooth device, etc.), so that the user can interact with the electronic device 500 via these external devices 100, and / or so that the electronic device 500 can communicate with one or more other data processing devices (such as a router, a modem, etc.). Such communication can be carried out through the input / output (I / O) interface 550, and can also be carried out with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 560. The network adapter 560 can communicate with other modules of the electronic device 500 through the bus 530. It should be understood that although Figure 5 not shown in the figure, other hardware and / or software modules can be used in the electronic device 500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID electronic devices, tape drives, and data backup storage electronic devices, etc.

[0201] Figure 6 is a schematic diagram of an embodiment of a computer-readable medium of the present invention. As Figure 6As shown, the computer program can be stored on one or more computer-readable media. The computer-readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electronic device, apparatus, or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. When the computer program is executed by one or more data processing devices, the computer-readable medium can implement the above method of the present invention, that is: classifying public opinion data into corresponding public opinion classification sets; performing dimensionality reduction processing on the public opinion data in the public opinion classification sets, and based on the first time window, searching for the historical event classification set to which the processed public opinion data belongs, and classifying the public opinion data into the historical event classification set to which it belongs; performing correlation matching on the public opinion data in the same historical event classification set, and classifying the matched public opinion data into the negative public opinion set; performing public opinion security analysis based on the public opinion classification set, the historical event classification set, and the negative public opinion set.

[0202] Through the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described in the present invention can be implemented by software or by a combination of software and necessary hardware. Therefore, the technical solution according to the embodiments of the present invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on the network, including several instructions to enable a data processing device (which can be a personal computer, a server, or a network device, etc.) to execute the above method according to the present invention.

[0203] The computer-readable storage medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, which can send, propagate, or transmit a program used by or in combination with an instruction execution electronic device, apparatus, or device. The program code contained on the readable storage medium can be transmitted by any appropriate medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination of the above.

[0204] The program code for performing the operations of the present invention can be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user's computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or, alternatively, can be connected to an external computing device (e.g., by connecting through the Internet using an Internet service provider).

[0205] In summary, the present invention can be implemented by a method, apparatus, electronic device, or computer-readable medium for executing a computer program. Some or all of the functions of the present invention can be implemented using a general-purpose data processing device such as a microprocessor or a digital signal processor (DSP) in practice.

[0206] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or electronic device, and various general-purpose devices can also implement the present invention. The above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A public opinion security analysis method based on event correlation, characterized in that, the method includes: Classify the public opinion data into corresponding public opinion classification sets; Convert the public opinion data in the public opinion classification set into binary, and split the converted string to obtain public opinion segments; Obtain each public opinion segment after dimensionality reduction and segmentation of the same public opinion text, obtain the string after dimensionality reduction processing of the public opinion text that occurred within the first time window, denoted as the string to be matched, perform correlation matching between each public opinion segment and each string to be matched to obtain the string to be compared; Compare the string where the public opinion segment is located with the string to be compared bit by bit. If the number of different bits is less than the predetermined value, use the string to be compared as the string similar to the public opinion data, search for the historical event classification set to which the string similar to the public opinion data belongs, use the historical event classification set with the most strings similar to the public opinion data as the historical event classification set to which the public opinion data belongs, and classify the public opinion data into the historical event classification set to which it belongs; Perform correlation matching on the public opinion data in the same historical event classification set, and classify the matched public opinion data into the negative public opinion set; Perform public opinion security analysis based on the public opinion classification set, historical event classification set, and the negative public opinion set.

2. The method according to claim 1, characterized in that, the method further includes: Perform dimensionality reduction processing on the newly added public opinion data, and based on the first time window, search for the historical event classification set to which the processed newly added public opinion data belongs, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event fermentation alarm message; If the historical event classification set to which the processed newly added public opinion data belongs cannot be found within the first time window, replace the first time window with the second time window, and based on the second time window, search for the historical event classification set to which the processed newly added public opinion data belongs, classify the newly added public opinion data into the historical event classification set to which it belongs, and send an event secondary fermentation alarm message; wherein: the second time window is earlier than the first time window.

3. The method according to claim 2, characterized in that, The comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the first time window is different from the comparison mechanism for searching for the historical event to which the processed newly added public opinion data belongs based on the second time window.

4. The method according to claim 1, characterized in that, the method further includes: Search and display the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set according to the user's search information.

5. The method according to claim 4, characterized in that, The searching and displaying the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set according to the user's search information includes: Search for the target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set based on multiple dimensions according to the user's search information; Determine the matching degree of each target public opinion according to the weight value of each dimension; Display each target public opinion according to the matching degree of the target public opinion.

6. The method according to claim 5, characterized in that, Before presenting each target public opinion according to the matching degree of the target public opinion, the method further includes: Adjusting the matching degree of the target public opinion according to the historical target public opinion matching the user's search information; or: Adjusting the matching degree of the target public opinion according to the final public opinion selected by the user from the historical target public opinion.

7. The method according to claim 5, wherein, The matching of the target public opinion based on multiple dimensions from the public opinion classification set, the historical event classification set, and the negative public opinion set according to the user's search information includes: Respectively matching the user's search information based on multiple dimensions from the public opinion classification set, the historical event classification set, and the negative public opinion set to obtain classified target public opinions, event target public opinions, and negative target public opinions.

8. The method according to claim 4, wherein, The method further includes: Constructing a user portrait according to the user's historical searches; Pushing information to the user based on the user portrait.

9. The method according to claim 1, wherein, The classifying and processing the public opinion data into the corresponding public opinion classification set includes: Performing a correlation analysis on the public opinion data to obtain relevant public opinion data; Classifying the relevant public opinion data according to its text attributes; Judging the sentiment type of the classified public opinion data based on the positive and negative sentiment corpus; Processing the sentiment-classified public opinion data using a classification model to obtain the public opinion classification set where the public opinion data is located.

10. The method according to claim 9, wherein, If the sentiment type of the classified public opinion data is negative, issuing a negative public opinion alert.

11. The method according to claim 9, wherein, The public opinion data includes: media public opinion data and social network public opinion data, and the method further includes: Performing a correlation matching on the social network public opinion data and classifying the matched social network public opinion data into the negative public opinion set.

12. An opinion security analysis device based on event correlation, wherein, The device includes: A classification module for classifying and processing public opinion data into the corresponding public opinion classification set; A dimensionality reduction and search module for converting the public opinion data in the public opinion classification set into binary, splitting the converted string to obtain public opinion segments; obtaining each public opinion segment after dimensionality reduction and splitting of the same public opinion text, obtaining the string after dimensionality reduction processing of the public opinion text occurring within the first time window, denoted as the string to be matched, performing a correlation matching on each public opinion segment and each string to be matched to obtain a string to be compared; comparing the string where the public opinion segment is located with the string to be compared bit by bit, if the number of different bits is less than a predetermined value, taking the string to be compared as the string similar to the public opinion data, searching for the historical event classification set to which the string similar to the public opinion data belongs, taking the historical event classification set containing the most strings similar to the public opinion data as the historical event classification set to which the public opinion data belongs, and classifying the public opinion data into the historical event classification set to which it belongs; A first matching module for performing a correlation matching on the public opinion data in the same historical event classification set and classifying the matched public opinion data into the negative public opinion set; An analysis module for performing public opinion security analysis based on the public opinion classification set, historical event classification set, and the negative public opinion set.

13. The apparatus according to claim 12, wherein, the apparatus further comprises: A first dimensionality reduction and search module for performing dimensionality reduction processing on newly added public opinion data, searching for the historical event classification set to which the processed newly added public opinion data belongs based on a first time window, classifying the newly added public opinion data into the historical event classification set to which it belongs, and sending an event fermentation alarm message; A second dimensionality reduction and search module for, if the historical event classification set to which the processed newly added public opinion data belongs is not found within the first time window, replacing the first time window with a second time window, searching for the historical event classification set to which the processed newly added public opinion data belongs based on the second time window, classifying the newly added public opinion data into the historical event classification set to which it belongs, and sending an event secondary fermentation alarm message; wherein: the second time window is earlier than the first time window.

14. The apparatus according to claim 13, wherein, the comparison mechanisms of the first dimensionality reduction and search module and the second dimensionality reduction and search module are different.

15. The apparatus according to claim 12, wherein, the apparatus further comprises: A search and display module for searching for and displaying target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set according to user search information.

16. The apparatus according to claim 15, wherein, the search and display module comprises: A multi-dimensional matching module for matching target public opinion from the public opinion classification set, historical event classification set, and negative public opinion set based on multiple dimensions according to user search information; A determination module for determining the matching degree of each target public opinion according to the weight value of each dimension; A display module for displaying each target public opinion according to the matching degree of the target public opinion.

17. The apparatus according to claim 16, wherein, the apparatus further comprises: An adjustment module for adjusting the matching degree of the target public opinion according to the historical target public opinion matching the user search information; or: adjusting the matching degree of the target public opinion according to the final public opinion selected by the user from the historical target public opinion.

18. The apparatus according to claim 16, wherein, the multi-dimensional matching module comprises: A sub-matching module for respectively matching user search information from the public opinion classification set, historical event classification set, and negative public opinion set based on multiple dimensions to obtain classified target public opinion, event target public opinion, and negative target public opinion.

19. The apparatus according to claim 15, wherein, the apparatus further comprises: A construction module for constructing a user portrait according to the user's historical search; A push module for pushing information to the user based on the user portrait.

20. The apparatus according to claim 12, wherein, the classification module comprises: A correlation analysis module for performing correlation analysis on public opinion data to obtain relevant public opinion data; A sub-classification module for classifying the relevant public opinion data according to the text attributes thereof; A judgment module, configured to judge the sentiment type of the classified public opinion data based on a positive and negative sentiment corpus; A model processing module, configured to process the classified public opinion data by using a classification model to obtain a public opinion classification set where the public opinion data is located.

21. The apparatus according to claim 20, wherein, the apparatus further includes: an alarm module, configured to send a negative public opinion alarm if the sentiment type of the classified public opinion data is negative.

22. The apparatus according to claim 20, wherein, the public opinion data includes: media public opinion data and social network public opinion data, and the apparatus further includes: A second matching module, configured to perform relevance matching on the social network public opinion data, and classify the matched social network public opinion data into a negative public opinion set.

23. An electronic device, including: a processor; and a memory storing computer-executable instructions, and when the computer-executable instructions are executed, the processor is caused to execute the method according to any one of claims 1-11.

24. A computer-readable storage medium, wherein, the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the method according to any one of claims 1-11 is implemented.

Images