Screening method, public opinion monitoring method and device for short video platform public opinion topics

Abstract

The present invention relates to the field of public opinion monitoring technology, and in particular to a method and device for screening public opinion topics on a short video platform; the method comprises: determining the monitoring field to which each topic monitoring account belongs; determining keywords used to trigger the screening of public opinion topics to form a first keyword set; monitoring the videos updated by each topic monitoring account; when any keyword in the first keyword set is included in a video updated by any topic monitoring account, marking the keyword as a first keyword; extracting the video containing the first keyword; extracting the videos with the first keyword that are published by the remaining topic monitoring accounts in the monitoring field within a first preset time, and identifying the text content in each video with the first keyword; analyzing and comparing the text content, and extracting the same character string as a public opinion topic. Determining public opinion topics based on video content effectively improves the accuracy of public opinion topics.

Description

Technical Field

[0001] The present invention relates to the field of welding technology, and in particular to a method for screening public opinion topics on a short video platform, a method for monitoring public opinion, and a device. Background Art

[0002] Public opinion is the sum of the beliefs, attitudes, opinions, and emotions expressed by the masses about various phenomena and problems in society. Online public opinion is a form of public opinion expression based on online media. It is formed more quickly, has a wider impact, and has a huge impact on society. With the rapid development of the Internet around the world, online media has been recognized as the "fourth media" after newspapers, radio, and television. With the development of science and technology and the accelerated pace of life, people have more and more outlets to express their opinions. In particular, the rapid rise of short video platforms has made it easier for people to understand various social events, and more and more people have flocked to the platform to spread the content of the event or express their views and attitudes on the event. However, as it becomes easier to express opinions, the opinions of keyboard warriors and others are becoming more and more likely to infect people. In order to understand the direction of people's opinions and public opinion in a timely manner, and to discover various problems reported by netizens as soon as possible, it becomes very necessary to use online public opinion as a clue and basis for solving problems. Through active response, we can make up for the shortcomings in our work or correct mistakes, so that the government or enterprises and institutions can develop better. At present, the main sources of public opinion information in the network environment are news websites, forums, blogs, Weibo, WeChat, etc. Many network media may generate public opinion, the monitoring scope is quite huge, and the monitoring workload is quite heavy. A major feature of network information is rapid dissemination, especially with the emergence of self-media such as Weibo and WeChat, the dissemination speed is growing geometrically. This requires us to discover relevant public opinions in real time, track and monitor in real time, and report and summarize in real time when doing monitoring. However, the time and energy of staff are limited, and manual detection is impossible. The current approach is to preset monitoring keywords according to the monitoring field, and the software compares whether there are exactly the same keywords in the hot topics on the Internet. When there are exactly the same keywords, it is determined that public opinion has been generated, and then the source of the public opinion is located. This reduces people's workload to a certain extent, but because each public opinion has its own characteristics, people cannot predict all the topic keywords, so the accuracy of its information is still very limited. When the monitored topic keywords are too simple or broad, too many events will be determined as public opinions, or even multiple completely unrelated niche events will be judged as a major public opinion event; and when the monitored topic is too precise or the semantic structure is too complex, certain events are easily missed; therefore, how to determine the effective topic keywords (referred to as topics) for an event has become an urgent problem that needs to be solved. Summary of the invention

[0003] The purpose of this application is to provide a method for screening public opinion topics on a short video platform, a public opinion monitoring method, a device and a storage medium. By setting up some topic monitoring accounts in the short video platform and setting keywords that match their fields, the videos updated by the topic monitoring accounts are analyzed. When the same keywords appear in multiple videos, the commonalities of the videos where the keywords appear are found as public opinion topics; the public opinion topics are determined based on the video content, effectively improving the accuracy of the public opinion topics, facilitating the rapid determination of public opinion events, and locating and obtaining public opinion content.

[0004] The present application provides a method for screening public opinion topics on a short video platform, the method comprising the following steps:

[0005] Classify the videos posted by the topic monitoring accounts according to their content, and determine the monitoring field to which each topic monitoring account belongs; determine the keywords used to trigger the screening of public opinion topics based on the monitoring field to form a first keyword set; monitor the videos updated by each topic monitoring account to determine whether the videos contain keywords in the first keyword set; when any video updated by any topic monitoring account contains any keyword in the first keyword set, mark the keyword as the first keyword; extract the video containing the first keyword, and identify the text content in the video; extract the videos containing the first keyword from the remaining videos posted by the topic monitoring accounts in the monitoring field within a first preset time, and identify the text content in each of the videos; analyze and compare the text content of the videos with the same first keyword, and extract the same character string as the public opinion topic.

[0006] Furthermore, the method for screening public opinion topics on a short video platform provided by the present application also includes: monitoring new comments on each short video of a topic monitoring account; when any new comment contains any keyword in the first keyword set, extracting the text content of the comment where the keyword is located, and marking the keyword as a second keyword; extracting videos with the second keyword from videos posted by the topic monitoring account in the monitoring field within a second preset time, and identifying the text content in each of the videos; analyzing and comparing the text content of videos with the same second keyword, and extracting the same character string as a public opinion topic.

[0007] Furthermore, topic monitoring accounts are classified according to the video content they post, and the monitoring field to which each topic monitoring account belongs is determined, specifically including: extracting short video accounts whose number of fans is greater than a first threshold as topic monitoring accounts; analyzing the video content posted by each topic monitoring account one by one, and determining the field to which each video content belongs; when a topic monitoring account has short videos greater than a second threshold in its video content that all belong to the same monitoring field, it is determined that the topic monitoring account belongs to the corresponding monitoring field.

[0008] Furthermore, the keywords used to trigger the screening of public opinion topics determined according to the monitoring field constitute a first keyword set, which specifically includes: extracting the characteristics of each topic that has generated public opinion in the monitoring field, as well as the commonly used keywords in the monitoring field, and recording them in the first keyword set; converting the keywords in the first keyword set into audio electrical signals in Mandarin and dialect forms.

[0009] Further, the monitoring of the videos updated by each topic monitoring account, determining whether there are keywords in the first keyword set in the video, and when any keyword in the first keyword set is included in the video updated by any topic monitoring account, marking the keyword as the first keyword, specifically includes:

[0010] Extract the audio from the videos of each topic monitoring account, and determine whether there is background music in the audio; if it is determined that there is background music, retrieve the music clip corresponding to the background music, and remove the corresponding music clip from the audio to obtain the audio to be identified; compare the electrical signal of the audio to be identified with the audio electrical signal in the first keyword set, extract the same content, convert it into text content, and obtain the first keyword corresponding to the audio of the topic monitoring account video; extract images in the video at preset time intervals; perform text recognition on the extracted images to obtain the text content of the image in the topic monitoring account video; compare the text content of the image with the content in the first keyword set, extract the same text content as in the first keyword set, and obtain the first keyword corresponding to the image in the topic monitoring account video.

[0011] Furthermore, when any keyword in the first keyword set is included in a video updated by any topic monitoring account, the keyword is marked as the first keyword, specifically including: when it is determined that any keyword in the first keyword set is included in a video updated by any topic monitoring account, the number of comments corresponding to the video is extracted; when the number of comments exceeds a preset comment threshold, it is determined that the keyword corresponding to the video updated by the topic monitoring account is the first keyword.

[0012] In another aspect of the present application, a method for monitoring public opinion on a short video platform is provided, comprising applying any of the above-described methods for screening public opinion topics on a short video platform to determine public opinion topics.

[0013] Furthermore, the public opinion monitoring method for the short video platform provided by the present invention also includes: collecting all comments and short videos with the topic in the short video platform.

[0014] On the other hand, the present application further provides a public opinion monitoring device for a short video platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the program.

[0015] In another aspect of the present application, a computer-readable storage medium is provided, which stores a computer program, wherein the program implements any of the above methods when executed by a processor.

[0016] The present application provides a method for screening public opinion topics on a short video platform, a method for monitoring public opinion, a device and a storage medium, which have the following improvements compared with the prior art:

[0017] 1. This application provides a method for screening public opinion topics on a short video platform. The method uses a topic monitoring account that can be used to screen public opinion topics based on the published videos and the monitoring field, sets a word with a wide coverage and a relatively simple structure as a keyword for the monitoring field and forms a first keyword set, then monitors the videos updated by each of the topic monitoring accounts, extracts the elements in the first keyword set that appear in the video as the first keyword, takes the time when the first keyword is extracted as the basis, searches for the videos updated by all the remaining topic monitoring accounts in the monitoring field within a certain period of time, and then searches for the common content in each video, thereby determining the design of the public opinion topic. By setting some topic monitoring accounts in the short video platform and setting keywords that match their fields, the videos updated by the topic monitoring accounts are analyzed. When the same keywords appear in multiple videos, the commonality of each video with the keywords is found as the public opinion topic. By using this method of associating keywords, topics that may form public opinion can be quickly found, and public opinion topics can be determined before public opinion is formed. It effectively solves the problem that topics are difficult to determine during public opinion monitoring in the prior art, and further solves the problem that inaccurate topic determination affects effective public opinion monitoring.

[0018] 2. The present application provides a method for screening public opinion topics on a short video platform, which uses the keywords determined according to the monitoring field to trigger the screening of public opinion topics to form a first keyword set, specifically including: extracting the characteristics of each topic that has generated public opinion in the monitoring field, as well as the commonly used keywords in the monitoring field, and recording them in the first keyword set; converting the keywords in the first keyword set into audio and electrical signals in the form of Mandarin and dialect. The first-level keyword set is composed of the characteristics of topics that have generated public opinion and the commonly used keywords in the monitoring field. The elements in the set are comprehensive, and the structure of each element is relatively simple, that is, it only needs to be a part of the topic that has formed public opinion; it does not have to be all of the topics that have formed public opinion. The text content in the first set is converted into audio and electrical signals in the form of Mandarin and dialect, which is convenient for comparing the audio and electrical signals when screening whether there are keywords in the audio of the video later, so as to quickly find out whether there are elements in the first keyword set. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0020] Figure 1 This is a flow chart of an embodiment of a method for screening public opinion topics on a short video platform provided by the present application;

[0021] Figure 2 It is a flowchart of a method for determining the monitoring field to which each topic monitoring account belongs in one embodiment of a method for screening public opinion topics on a short video platform provided by the present application;

[0022] Figure 3 It is a flowchart of determining the first keyword in an embodiment of a method for screening public opinion topics on a short video platform provided by the present application. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0024] Figure 1It is a method flow chart of an embodiment of a method for screening public opinion topics on a short video platform described in the present application. Although the present application provides method operation steps as shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order described in the embodiments of the present application and shown in the drawings. When the method is applied in an actual device or terminal product, it can be connected in sequence or in parallel according to the method shown in the embodiments or drawings (for example, a parallel processor or a multi-threaded processing environment, or even a distributed processing environment).

[0025] The present application provides a method for screening public opinion topics on a short video platform. The method uses a big V short video account as a topic monitoring account, classifies it, determines the field it belongs to, and then determines the first keyword set corresponding to the video of the short video account; then analyzes the videos updated by each topic monitoring account, extracts each video with elements in the first keyword set, compares each video, and extracts the same character strings as public opinion topics; when adopting the method, each element in the first keyword set can be a relatively simple word or a single character, and of course it can also be a company name, etc., which is equivalent to first using the first keyword for a rough screening, and then using the content in the video for fine selection. Through multiple analysis and extraction, it is effectively realized to accurately determine the public opinion topic, thereby ensuring the effectiveness, accuracy and timeliness of public opinion monitoring.

[0026] Specific as Figure 1 As shown, a method for screening public opinion topics on a short video platform provided by an embodiment of the present application includes the following steps:

[0027] S1: Classify the videos posted by topic monitoring accounts according to their content, and determine the monitoring field to which each topic monitoring account belongs;

[0028] S2: Determine keywords for triggering public opinion topic screening according to the monitoring field to form a first keyword set;

[0029] S3: monitor the videos updated by each topic monitoring account, and determine whether there are keywords in the first keyword set in the video; when any keyword in the first keyword set is included in a video updated by any topic monitoring account, mark the keyword as the first keyword;

[0030] S4: extracting the video containing the first keyword and identifying the text content in the video;

[0031] S5: extracting videos containing the first keyword from the videos posted by the remaining topic monitoring accounts in the monitoring field within a first preset time, and identifying text content in each of the videos;

[0032] S6: Analyze and compare the text contents of the videos having the same first keyword, and extract the same character string as the public opinion topic.

[0033] In this embodiment, the text content of the videos with the same first keyword is analyzed and compared, and the same character string is extracted as the public opinion topic, specifically:

[0034] Extract N times (usually N=2) the character length of the first keyword as the reference character length;

[0035] Extract 2 to the benchmark character length of consecutive characters from the text content of the first video as a benchmark character string;

[0036] Comparing each benchmark character string with the text content of each video having the first keyword in the remaining topic monitoring accounts in the monitoring field;

[0037] In the text content of the remaining topic monitoring account videos, each additional occurrence of the benchmark string is counted as one occurrence;

[0038] Count the number of times each benchmark string appears again, and sort them from most to least according to the number of occurrences;

[0039] When the number of occurrences of the first character string is much greater than the number of occurrences of the second character string, the first character string is determined to be the public opinion topic;

[0040] When the number of occurrences of the string in the first position is close to the number of occurrences of the string in the second position, and the number of occurrences of the string in the second position is much greater than the number of occurrences of the string in the third position, the similarity between the string in the first position and the string in the second position is judged, and when one of the strings is contained in the other, the string with the larger character length is extracted as the public opinion topic; when the two strings do not belong to a contained-in relationship, it is determined that the string in the first position and the string in the second position are both the public opinion topics.

[0041] The present application provides a method for screening public opinion topics on a short video platform, which adopts a method for determining a topic monitoring account that can be used to screen public opinion topics based on the published videos and the monitoring field, setting a word with a wide coverage and a relatively simple structure as a keyword for the monitoring field and forming a first keyword set, and then monitoring the videos updated by each of the topic monitoring accounts, extracting the elements in the first keyword set that appear in the video as the first keyword, taking the time when the first keyword is extracted as the basis, searching for the videos updated by all the remaining topic monitoring accounts in the monitoring field within a certain period of time, and then searching for the common content in each video, so as to determine the design of the public opinion topic. By setting some topic monitoring accounts in the short video platform and setting keywords that match their fields, analyzing the videos updated by the topic monitoring accounts, when the same keywords appear in multiple videos, finding the commonality of each video where the keywords appear as the public opinion topic. Using this method of associating keywords, you can quickly find topics that may form public opinion, and determine the public opinion topic before the public opinion is formed. It effectively solves the problem that topics are not easy to determine in public opinion monitoring in the prior art, and then solves the problem that topics are not accurately determined, thereby affecting the effective monitoring of public opinion.

[0042] In this embodiment, steps S1 and S2 are interchangeable. You can first screen the video accounts in the short video platform to obtain topic monitoring accounts, or you can first determine the keywords of the monitoring field. The keywords of each monitoring field may be different or may overlap. A short video account can be used as a topic monitoring account for only one monitoring field, or it can be a topic monitoring account for multiple monitoring fields.

[0043] The method for screening public opinion topics on the short video platform provided in this embodiment also includes:

[0044] Monitor new comments on each short video of the topic monitoring account;

[0045] When any newly added comment contains any keyword in the first keyword set, extract the text content of the comment containing the keyword, and mark the keyword as a second keyword;

[0046] Extracting videos containing the second keyword from videos posted by the topic monitoring account in the monitoring field within a second preset time, and identifying text content in each of the videos;

[0047] The text content of the videos having the same second keyword is analyzed and compared, and the same character string is extracted as a public opinion topic. The length of the same character string is greater than 3.

[0048] In this embodiment, the review content is monitored by using OCR text recognition technology, and the characters in the image are recognized by using OCR text recognition. This is a prior art and will not be described in detail here.

[0049] like Figure 2 As shown, in this embodiment, the video content released by the topic monitoring account is classified to determine the monitoring field to which each topic monitoring account belongs, specifically including:

[0050] S11: extracting short video accounts whose number of fans is greater than a first threshold as topic monitoring accounts;

[0051] S12: Analyze the video content posted by each topic monitoring account one by one, and determine the field to which each video content belongs;

[0052] S13: When the number of short videos greater than a second threshold in the video content of a topic monitoring account all belongs to the same monitoring field, it is determined that the topic monitoring account belongs to the corresponding monitoring field.

[0053] In this embodiment, the first threshold can be 10,000 to 500,000, which can be determined specifically according to the degree of attention people pay to the monitored field. The top 5 to 15 short video accounts with a large number of fans on the platform can also be selected as the topic monitoring accounts. The second threshold can be any value from 55% to 80%. When the monitored field is a relatively popular field or a field with a large number of followers, a larger value can be selected, such as in the financial field, 80% can be selected; that is, when the number of videos in the financial field accounts for more than 80% of the total number of videos released, the short video account is determined to belong to the financial field.

[0054] In this embodiment, the keywords used to trigger the screening of public opinion topics determined according to the monitoring field constitute a first keyword set, which specifically includes:

[0055] Extracting the features of various topics that have generated public opinions in the monitoring field, as well as common keywords in the monitoring field, and recording them in the first keyword set;

[0056] The keywords in the first keyword set are converted into audio electrical signals in Mandarin and dialect.

[0057] The audio electrical signal actually refers to the current that changes in the circuit according to the audio law. In the public opinion topics that have appeared in the monitoring field in the past, words that appear more than twice can be used as the characteristics of each topic. Common keywords in this field can be company names, names, positions, industry terms, place names, organization names, etc. The converted audio electrical signals can be stored in the form of Mandarin packages and first dialect packages, second dialect packages, and Nth dialect packages.

[0058] The present application provides a method for screening public opinion topics on a short video platform, which uses the keywords determined according to the monitoring field to trigger the screening of public opinion topics to form a first keyword set, specifically including: extracting the characteristics of each topic that has generated public opinion in the monitoring field, as well as the commonly used keywords in the monitoring field, and recording them in the first keyword set; converting the keywords in the first keyword set into audio and electrical signals in the form of Mandarin and dialect. The first-level keyword set is composed of the characteristics of topics that have generated public opinion and the commonly used keywords in the monitoring field. The elements in the set are comprehensive, and the structure of each element is relatively simple, that is, it only needs to be a part of the topic that has formed public opinion; it does not have to be all of the topics that have formed public opinion. The text content in the first set is converted into audio and electrical signals in the form of Mandarin and dialect, which is convenient for comparing the audio and electrical signals when screening whether there are keywords in the audio of the video later, so as to quickly find out whether there are elements in the first keyword set.

[0059] Since short videos usually have background music, directly performing speech recognition will affect the content of the recognized text; at the same time, if speech recognition is performed on the audio of each short video, a large amount of calculation will be generated, requiring a lot of computing resources, and will also cause a long time problem. In order to reduce the amount of calculation and increase the corresponding rate of obtaining the first keyword while ensuring that the first keyword can be obtained, this embodiment adopts the following method to obtain the first keyword.

[0060] like Figure 3 In this embodiment, the process of monitoring the videos updated by each topic monitoring account, determining whether there are keywords in the first keyword set in the video, and marking the keyword as the first keyword when any of the keywords in the first keyword set are found in the videos updated by any topic monitoring account, specifically includes:

[0061] S31: extracting the audio in the video of each topic monitoring account, and determining whether there is background music in the audio;

[0062] S32: if it is determined that there is background music, retrieve a music clip corresponding to the background music, and remove the corresponding music clip from the audio to obtain the audio to be recognized;

[0063] S33: Compare the electric signal of the audio to be identified with the audio electric signal in the first keyword set, extract the same content, convert it into text content, and obtain the first keyword corresponding to the audio of the video of the topic monitoring account;

[0064] S34: extracting images from the video at preset time intervals;

[0065] S35: Perform text recognition on the extracted image to obtain the text content of the image in the video of the topic monitoring account;

[0066] S36: Compare the text content of the image with the content in the first keyword set respectively, extract the same text content as in the first keyword set, and obtain the first keyword corresponding to the image in the video of the topic monitoring account.

[0067] In this embodiment, the order of S31-S33 and S34-S35 can be reversed.

[0068] In this embodiment, comparing the electric signal of the audio to be recognized with the electric signal of the audio in the first keyword set includes:

[0069] Determine whether the language of the audio to be recognized is a dialect or Mandarin, and determine a corresponding audio electrical signal package;

[0070] The corresponding audio electrical signal packet is extracted and compared with the electrical signal of the audio to be identified.

[0071] This embodiment adopts the method of directly comparing the audio to search for the first keyword in the audio, which effectively reduces the workload of voice recognition. At the same time, in order to avoid missed detection, image recognition is also performed on the video, and the contents of image recognition and voice recognition are respectively compared with the first keyword set. The union of the first keywords obtained effectively ensures the comprehensiveness of obtaining the first keyword.

[0072] This embodiment uses images and music to achieve no missed detection; extracting images at preset time intervals reduces the amount of data processing, improves the data processing speed, reduces repeated detection content, and is conducive to quickly detecting keywords in the video. Furthermore, by analyzing the updated video style of the topic monitoring account, determining the commonly used background images of the topic monitoring account, and removing the commonly used background images during image recognition, it can effectively reduce the amount of calculation for character recognition in the image, improve the speed of determining public opinion topics, and quickly grasp the public opinion situation.

[0073] In this embodiment, when any keyword in the first keyword set is included in a video updated by any topic monitoring account, marking the keyword as the first keyword specifically includes:

[0074] When it is determined that a video updated by any topic monitoring account contains any keyword in the first keyword set, extracting the number of comments corresponding to the video;

[0075] When the number of comments exceeds a preset comment threshold, it is determined that the keyword corresponding to the video updated by the topic monitoring account is the first keyword.

[0076] In this embodiment, the preset comment threshold is 100 to 5000. It can be selected according to actual needs. When determining the specific value, it can also be determined by referring to the field in which the topic monitoring account is located and the number of fans of the topic monitoring account. The more people pay attention to the field and the more fans there are, the larger the value of the preset comment threshold can be set.

[0077] This embodiment screens the preliminarily determined keywords by the number of comments, reduces the possibility of monitoring topics that are not generated by public opinion, and improves the accuracy of public opinion topics.

[0078] Based on the same inventive concept, an embodiment of the present application also provides a method for monitoring public opinion on a short video platform, comprising the following steps:

[0079] Apply the above-mentioned short video platform public opinion topic screening method to determine the public opinion topic;

[0080] Collect all comments and short videos on the short video platform with the above topic.

[0081] You can also perform secondary sorting on the public opinion data on the mobile phone, such as descending time order, ascending time order, popularity order, etc.; you can also perform emotional screening, such as positive, negative, and neutral.

[0082] Based on the same inventive concept, an embodiment of the present application also provides a public opinion monitoring device for a short video platform, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in any of the above embodiments when executing the program. Since the principle of solving the problem by the public opinion monitoring device for a short video platform is similar to that of the public opinion monitoring method for a short video platform, the implementation of the public opinion monitoring device for a short video platform can refer to the implementation of the public opinion monitoring method for a short video platform, and the repeated parts will not be repeated.

[0083] Based on the same inventive concept, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which implements the method described above when executed by a processor.

[0084] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary hardware. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product or in the implementation process of data migration. All or part of the present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, mobile communication terminals, multi-processor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0085] Although the present application has been described through embodiments, persons skilled in the art will appreciate that there are many modifications and variations of the present application without departing from the spirit of the present application. It is intended that the appended claims include these modifications and variations without departing from the spirit of the present application.

Claims

1. A method for screening public opinion topics on a short video platform, characterized in that: The method comprises the following steps: Classify the videos posted by topic monitoring accounts according to their content, and determine the monitoring field to which each topic monitoring account belongs; Determine keywords for triggering public opinion topic screening according to the monitoring field to form a first keyword set; Monitor the videos updated by each topic monitoring account, and determine whether there are keywords in the first keyword set in the video; when any keyword in the first keyword set is included in a video updated by any topic monitoring account, mark the keyword as the first keyword; Extracting the video containing the first keyword and identifying the text content in the video; Extracting videos containing the first keyword from the videos posted by the remaining topic monitoring accounts in the monitoring field within a first preset time, and identifying text content in each of the videos; Analyze and compare the text content of videos with the same first keyword, and extract the same character string as the public opinion topic; The step of monitoring the videos updated by each topic monitoring account, determining whether there are keywords in the first keyword set in the videos, and marking the keywords as the first keywords when any of the keywords in the first keyword set are included in the videos updated by any topic monitoring account, specifically includes: Extracting the audio from the video of each topic monitoring account, and determining whether there is background music in the audio; When it is determined that there is background music, a music clip corresponding to the background music is retrieved, and the corresponding music clip is removed from the audio to obtain the audio to be recognized; Compare the electric signal of the audio to be identified with the audio electric signal in the first keyword set, extract the same content, convert it into text content, and obtain the first keyword corresponding to the audio of the video of the topic monitoring account; Extracting images from the video at preset time intervals; Perform text recognition on the extracted image to obtain the text content of the image in the video of the topic monitoring account; The text content of the image is compared with the content in the first keyword set respectively, and the text content identical to that in the first keyword set is extracted to obtain the first keyword corresponding to the image in the video of the topic monitoring account.

2. The method according to claim 1, characterized in that Also includes: Monitor new comments on each short video of the topic monitoring account; When any newly added comment contains any keyword in the first keyword set, extract the text content of the comment containing the keyword, and mark the keyword as a second keyword; Extracting videos containing the second keyword from videos posted by the topic monitoring account in the monitoring field within a second preset time, and identifying text content in each of the videos; The text contents of the videos having the same second keyword are analyzed and compared, and the same character strings are extracted as public opinion topics.

3. The method according to claim 2, characterized in that The topic monitoring accounts are classified according to the video content they publish, and the monitoring field to which each topic monitoring account belongs is determined, specifically including: Extract short video accounts with a follower count greater than a first threshold as topic monitoring accounts; Analyze the video content posted by each topic monitoring account one by one to determine the field to which each video content belongs; When the short videos with a number greater than a second threshold in the video content of a topic monitoring account all belong to the same monitoring field, it is determined that the topic monitoring account belongs to the corresponding monitoring field.

4. The method according to claim 3, characterized in that The keywords used to trigger the screening of public opinion topics determined according to the monitoring field constitute a first keyword set, which specifically includes: Extracting the features of various topics that have generated public opinions in the monitoring field, as well as common keywords in the monitoring field, and recording them in the first keyword set; The keywords in the first keyword set are converted into audio electrical signals in Mandarin and dialect.

5. The method according to claim 4, characterized in that When any keyword in the first keyword set is included in a video updated by any topic monitoring account, marking the keyword as the first keyword specifically includes: When it is determined that a video updated by any topic monitoring account contains any keyword in the first keyword set, extracting the number of comments corresponding to the video; When the number of comments exceeds a preset comment threshold, it is determined that the keyword corresponding to the video updated by the topic monitoring account is the first keyword.

6. A method for monitoring public opinion on a short video platform, characterized in that: The public opinion topic is determined by applying the method for screening public opinion topics on a short video platform as described in any one of claims 1 to 5.

7. The method according to claim 6, characterized in that Collect all comments and short videos on the short video platform with the above topic.

8. A public opinion monitoring device for a short video platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images