A method and system for predicting public opinion development trend

Abstract

The present invention belongs to the technical field of public opinion management. A method and system for predicting the development trend of public opinion are provided. The method comprises: obtaining the public opinion big data of the target public opinion to be analyzed, using the event clustering tracking module to extract and cluster the public opinion big data, and obtaining the initial event set; obtaining the incremental data of the public opinion big data of the target public opinion according to the backtracking period, and the event clustering tracking module extracts and clusters the incremental data to obtain a number of newly added event clusters; matching and analyzing each newly added event cluster with each event cluster in the initial event set, if the match is successful, the newly added event cluster is merged into the corresponding event cluster, otherwise the newly added event cluster is added to the initial event set to obtain a dynamic event set; the initial event set or the dynamic event set is input into the public opinion trend analysis model based on the LLM model to obtain the public opinion development trend report of the target public opinion. The present invention realizes the accurate prediction and analysis of the development trend of public opinion.

Description

Technical Field

[0001] The present invention relates to the technical field of public opinion management, and in particular to a method and system for predicting the development trend of public opinion. Background Art

[0002] In today's information-based and networked era, the power of public opinion is becoming increasingly powerful, and public opinion analysis has become the focus of attention of governments, enterprises, media and other parties. However, the analysis module in traditional public opinion products still mainly relies on regular analysis generated by regular back-end calculations, lacking specific analysis of individual events, such as analysis of the trend of public opinion, which makes it impossible to guide relevant personnel to quickly make the best response measures to public opinion in a timely manner. The present invention aims to achieve accurate prediction of the development trend of public opinion. Summary of the invention

[0003] In this regard, the present invention provides a method, system, electronic device, computer storage medium and computer program product for predicting the development trend of public opinion to solve the above-mentioned technical problems.

[0004] The present invention discloses a method for predicting the development trend of public opinion, which comprises the following steps: obtaining public opinion big data of a target public opinion to be analyzed, using an event clustering tracking module to perform event extraction and clustering processing on the public opinion big data, and obtaining an initial event set; wherein the initial event set contains a number of event clusters; obtaining incremental data of the public opinion big data of the target public opinion according to a backtracking period, and the event clustering tracking module performs event extraction and clustering processing on the incremental data, and obtains a number of newly added event clusters; matching and analyzing each of the newly added event clusters with each of the event clusters in the initial event set, if the match is successful, merging the newly added event cluster into the corresponding event cluster, otherwise supplementing the newly added event cluster into the initial event set, and obtaining a dynamic event set; wherein both the event cluster and the newly added event cluster contain event intensity values; inputting the initial event set or the dynamic event set into a public opinion trend analysis model based on an LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein the public opinion development trend report contains a public opinion development trend segment after the current moment.

[0005] In some embodiments, the public opinion trend analysis model is trained in the following manner: collecting model training data; wherein the model training data includes public opinion industry reports, public opinion analysis cases, and public opinion analyst summary data; using the model training data to train the public opinion trend analysis model, and publishing the trained public opinion trend analysis model.

[0006] In some embodiments, the event clustering tracking module is used to extract and cluster the public opinion big data to obtain an initial event set, including: the event clustering tracking module calls an AI engine to implement: extracting event element information from the public opinion big data, the event element information includes location, person, and event type, and abstractly summarizing the location, the person, and the event type to obtain event element summary information; using a clustering algorithm to cluster each of the event element summary information to achieve event classification, and obtain a number of event clusters; and, performing statistical analysis on each of the event element information corresponding to each of the event clusters to obtain the event intensity value; after associating each of the event clusters with the corresponding event intensity value, integrating them to obtain the initial event set.

[0007] In some embodiments, the obtaining of the incremental data of the public opinion big data of the target public opinion according to the backtracking period includes: using the public opinion trend analysis model to perform a shallow analysis on the initial event set to obtain the public opinion type of the target public opinion, and determining a first backtracking period based on the public opinion type; calculating the number of event clusters in the initial event set or the dynamic event set, calculating a backtracking adjustment coefficient based on the number of event clusters, multiplying the backtracking adjustment coefficient by the first backtracking period to obtain a second backtracking period, which is the backtracking period; obtaining the incremental data of the public opinion big data of the target public opinion according to the backtracking period.

[0008] In some embodiments, the initial event set or the dynamic event set is input into a public opinion trend analysis model based on an LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion, including: inputting the initial event set or the dynamic event set and manually specified predicted duration data into a public opinion trend analysis model based on an LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the duration of the public opinion development trend segment after the current moment in the public opinion development trend report matches the predicted duration data.

[0009] In some embodiments, the method further includes: analyzing and obtaining the activity corresponding to each event cluster based on the event intensity value of each event cluster in the dynamic event set, and clearing the event clusters whose activity is lower than an activity threshold.

[0010] The present invention also discloses a public opinion development trend prediction system, which includes a first event extraction module, a second event extraction module, an event set dynamic update module, and a public opinion trend prediction module; the first event extraction module is used to obtain the public opinion big data of the target public opinion to be analyzed, and use the event clustering tracking module to perform event extraction and clustering processing on the public opinion big data to obtain an initial event set; wherein the initial event set contains a number of event clusters; the second event extraction module is used to obtain incremental data of the public opinion big data of the target public opinion according to the backtracking period, and the event clustering tracking module performs event extraction and clustering processing on the incremental data to obtain a number of newly added event clusters; the event set dynamic update module is used to compare each of the newly added event clusters with each event in the initial event set. The event cluster is matched and analyzed. If the match is successful, the newly added event cluster is merged into the corresponding event cluster, otherwise the newly added event cluster is added to the initial event set to obtain a dynamic event set; wherein, both the event cluster and the newly added event cluster contain event intensity values; the public opinion trend prediction module is used to input the initial event set or the dynamic event set into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the public opinion development trend report contains the public opinion development trend segment after the current moment.

[0011] The present invention also discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method described in any of the preceding items.

[0012] The present invention also discloses a computer storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement any of the above methods.

[0013] The present invention also discloses a computer program product. When the computer program product is run on a terminal, the terminal implements any of the above methods when executing the computer program product.

[0014] The beneficial effects of the present invention are as follows: the scheme of the present invention analyzes the development trend of public opinion based on various events in public opinion, and constructs a public opinion trend analysis model based on the LLM model to perform trend analysis on the sorted event set, thereby obtaining a final public opinion development trend report, wherein the report predicts the public opinion trend for the next period of time based on the historical development trend of the target public opinion, which can assist relevant personnel in evaluating the development of the target public opinion, thereby formulating reasonable response measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flow chart of a method for predicting the development trend of public opinion disclosed in an embodiment of the present invention.

[0017] Figure 2 It is a schematic diagram of generating an initial event set and a dynamic event set disclosed in an embodiment of the present invention.

[0018] Figure 3 It is a structural diagram of a public opinion development trend prediction system disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following is a description of the implementation of the present application by specific specific embodiments. People familiar with the technology can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0020] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0021] like Figure 1 , 2As shown, an embodiment of the present invention discloses a method for predicting the development trend of public opinion, which includes the following steps: obtaining public opinion big data of the target public opinion to be analyzed, using an event clustering tracking module to perform event extraction and clustering processing on the public opinion big data, and obtaining an initial event set; wherein the initial event set contains a number of event clusters; obtaining incremental data of the public opinion big data of the target public opinion according to the backtracking period, and the event clustering tracking module performs event extraction and clustering processing on the incremental data to obtain a number of newly added event clusters; matching and analyzing each of the newly added event clusters with each of the event clusters in the initial event set, if the match is successful, merging the newly added event cluster into the corresponding event cluster, otherwise adding the newly added event cluster into the initial event set to obtain a dynamic event set; wherein both the event cluster and the newly added event cluster contain event intensity values; inputting the initial event set or the dynamic event set into a public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein the public opinion development trend report contains the public opinion development trend segment after the current moment.

[0022] The solution of the present invention analyzes the development trend of public opinion based on various events in public opinion, and constructs a public opinion trend analysis model based on the LLM model to perform trend analysis on the sorted event set, thereby obtaining a final public opinion development trend report, wherein the report predicts the public opinion trend for the next period of time based on the historical development trend of the target public opinion, which can assist relevant personnel in evaluating the development of the target public opinion and thus formulate reasonable response measures.

[0023] Specifically, first obtain (for example, through web crawler technology) the public opinion big data of the target public opinion to be analyzed, use the pre-built event clustering tracking module to extract and cluster the events, and obtain an initial event set containing several event clusters; then, periodically obtain the public opinion big data according to the backtracking period to obtain the corresponding new data, use the event clustering tracking module to perform the same processing on the new data, obtain several new event clusters, match and analyze them with each event cluster in the initial event set, if the match is successful, merge the new event cluster with the successfully matched event cluster, otherwise it is identified as a new event and added to the initial event set. After multiple acquisitions and analyses of the new data, a dynamic event set that can be dynamically updated is obtained. Finally, the dynamic event set of the public opinion trend analysis model based on the LLM model is used for in-depth analysis, and a public opinion development trend report based on the current moment can be obtained.

[0024] Among them, the event clusters in the initial event set and the dynamic event set (including newly added event clusters) contain event codes, which are used to identify a specific event related to the target public opinion. In addition, it also includes the corresponding event intensity value, which is used to describe the "heat" of a specific event, and the "heat" is composed of the number of topics, the number of participating users, the scope of involvement (region, field, etc.), the number of searches, etc. The event intensity value can be used in the public opinion trend analysis model to understand the development trend of the target public opinion, so as to generate the corresponding public opinion development trend report.

[0025] In some embodiments, the public opinion trend analysis model is trained in the following manner: collecting model training data; wherein the model training data includes public opinion industry reports, public opinion analysis cases, and public opinion analyst summary data; using the model training data to train the public opinion trend analysis model, and publishing the trained public opinion trend analysis model.

[0026] In the embodiment of the present invention, the LLM model is a large language model, which is a natural language processing technology based on deep learning. By learning massive text data, it can understand the meaning, context and logical relationship of language, thereby achieving tasks such as natural language generation, text classification, and sentiment analysis. Therefore, the present invention applies the LLM model to public opinion analysis and constructs a public opinion trend analysis model based on the LLM model.

[0027] Before publishing the above-mentioned public opinion trend analysis model, the present invention first collects three types of data, including public opinion industry reports, public opinion analysis cases, and public opinion analyst summary data, which include public opinion analysis dimensions, public opinion event development laws, report writing methods and other analysis methods, and constitutes model training data. The model training data constructed by combining the above three types of data can fully train the public opinion trend analysis model, and the trained public opinion trend analysis model can automatically organize, analyze and generate reports on online speeches.

[0028] It should be noted that the above-mentioned public opinion development trend report generated by the public opinion trend analysis model itself can also be used as model training data to achieve update training of its own model.

[0029] In some embodiments, the event clustering tracking module is used to extract and cluster the public opinion big data to obtain an initial event set, including: the event clustering tracking module calls an AI engine to implement: extracting event element information from the public opinion big data, the event element information includes location, person, and event type, and abstractly summarizing the location, the person, and the event type to obtain event element summary information; using a clustering algorithm to cluster each of the event element summary information to achieve event classification, and obtain a number of event clusters; and, performing statistical analysis on each of the event element information corresponding to each of the event clusters to obtain the event intensity value; after associating each of the event clusters with the corresponding event intensity value, integrating them to obtain the initial event set.

[0030] In an embodiment of the present invention, the event clustering tracking module includes an AI engine, which can be used to extract and cluster the acquired public opinion big data to obtain an initial event set. Specifically, the AI ​​engine first extracts event element information from the public opinion big data, and the event element information includes location, person, and event type. Then the location, person, and event type are abstractly summarized to obtain event element summary information, for example, the person is abstractly summarized as workers, students, public officials, merchants, etc., and the location is abstractly summarized as residences, supermarkets, transportation hubs, campuses, etc. Then, the clustering algorithm is used to cluster the obtained event element summary information, that is, multiple sub-events with the same attributes are clustered together (for example, those belonging to the same transportation hub are clustered together) to form an event cluster, thereby completing the classification of the event. The AI ​​engine also performs statistical analysis on the event element information corresponding to each event cluster to obtain the corresponding event intensity value, and the specific analysis method is the aforementioned "heat" determination method. After the calculated event intensity value is associated with the corresponding event cluster, all event clusters constitute the initial event set.

[0031] In some embodiments, the obtaining of the incremental data of the public opinion big data of the target public opinion according to the backtracking period includes: using the public opinion trend analysis model to perform a shallow analysis on the initial event set to obtain the public opinion type of the target public opinion, and determining a first backtracking period based on the public opinion type; calculating the number of event clusters in the initial event set or the dynamic event set, calculating a backtracking adjustment coefficient based on the number of event clusters, multiplying the backtracking adjustment coefficient by the first backtracking period to obtain a second backtracking period, which is the backtracking period; obtaining the incremental data of the public opinion big data of the target public opinion according to the backtracking period.

[0032] In an embodiment of the present invention, the event clustering tracking module in the present invention can realize continuous updating of the dynamic event set by continuously acquiring the public opinion big data of the target public opinion, that is, completing the continuous updating and heat tracking of the event. However, if the above tracking is achieved by performing an overall analysis of the new and old public opinion big data every time, it will lead to a huge waste of computing power and will also cause the timeliness of the tracking to deteriorate. In this regard, the present invention is configured to obtain the incremental data of the public opinion big data of the target public opinion according to the backtracking period, and only perform dynamic analysis of the event on the incremental data.

[0033] The backtracking period is dynamic, and its determination method is as follows: First, use the public opinion trend analysis model to perform a shallow analysis on the initial event set to obtain the public opinion type of the target public opinion. It should be noted here that the public opinion trend analysis model is configured with a variety of analysis modes, such as a deep analysis mode and a shallow analysis mode. The deep analysis mode implements a deep analysis of the event set to form a complete public opinion development trend report, while the shallow analysis mode only analyzes and processes certain specific data in the public opinion development trend report and outputs it. This embodiment uses the shallow analysis mode to obtain the public opinion type to which the target public opinion obtained by the public opinion trend analysis model belongs. Among them, the public opinion type can be conventional news public opinion, major malicious incident public opinion, etc. The above is only an example, and is not used to limit the public opinion type in the present invention to only include the above-mentioned division method. In fact, it can be divided into multiple different levels according to the duration of the public opinion.

[0034] After determining the type of the target public opinion, the first tracing period can be determined according to the public opinion type. Different public opinion types are pre-configured with different first tracing periods. For example, the first tracing period for regular news public opinion is 7 days, while the first tracing period for major malicious incidents is 15 or 30 days. Obviously, the higher the level, the longer the corresponding first tracing period.

[0035] Then, the number of event clusters in the initial event set or dynamic event set is statistically calculated, and the retrospective adjustment coefficient is calculated based on the number of event clusters. The retrospective adjustment coefficient is multiplied by the first retrospective period obtained above to obtain the final retrospective period. The more event clusters there are in the initial event set or dynamic event set, the higher the complexity of the target public opinion. At this time, a larger retrospective adjustment coefficient should be set to appropriately increase the first retrospective period obtained above, and the accuracy of the newly added event clusters obtained based on more incremental data analysis is higher; conversely, the fewer event clusters there are in the initial event set or dynamic event set, the lower the complexity of the target public opinion. At this time, a smaller retrospective adjustment coefficient should be set to appropriately reduce the first retrospective period obtained above, and the newly added event clusters are analyzed based on less incremental data, which can ensure that the newly added event clusters obtained by analysis have higher accuracy, and the analysis efficiency is also higher due to less incremental data used for analysis.

[0036] In some embodiments, the initial event set or the dynamic event set is input into a public opinion trend analysis model based on an LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion, including: inputting the initial event set or the dynamic event set and manually specified predicted duration data into a public opinion trend analysis model based on an LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the duration of the public opinion development trend segment after the current moment in the public opinion development trend report matches the predicted duration data.

[0037] In an embodiment of the present invention, the public opinion trend analysis model can analyze the aforementioned determined initial event set or each dynamic event set, so as to generate the above-mentioned public opinion development trend report, which is obtained by comprehensively analyzing the distribution of event clusters in public opinion and event intensity values. The specific display form of the report can be textual form, trend graph form, a mixture of the two or other forms, and the present invention does not limit this. At the same time, based on the different needs of relevant personnel, the forecast duration data can also be manually specified. The forecast duration data is used to indicate the time span of the "forecast trend segment" of the public opinion development trend report generated by the public opinion trend analysis model, such as 3 days or 5 days, that is, the last segment of the development trend graph in the public opinion development trend report that the relevant personnel need to generate contains the public opinion trend of the above-mentioned specified forecast duration, so as to determine the response strategy.

[0038] In some embodiments, the method further includes: analyzing and obtaining the activity corresponding to each event cluster based on the event intensity value of each event cluster in the dynamic event set, and clearing the event clusters whose activity is lower than an activity threshold.

[0039] In an embodiment of the present invention, in the process of dynamically updating the dynamic event set, the event intensity values ​​corresponding to each event cluster (including newly added event cluster) contained therein are also subjected to trend analysis to obtain the corresponding activity. The activity is calculated based on the latest event intensity value and the downward / increasing trend. The smaller the latest event intensity value and the greater the downward trend, the lowest activity; the larger the latest event intensity value and the greater the upward trend, the highest activity; and other situations are intermediate situations. Different activity values ​​are configured accordingly.

[0040] If the activity of an event cluster is lower than the activity threshold, it means that it is no longer in the development trend of the target public opinion. At this time, it will be cleaned up to reduce storage and processing requirements.

[0041] like Figure 3As shown, an embodiment of the present invention further discloses a public opinion development trend prediction system, the system comprising a first event extraction module, a second event extraction module, an event set dynamic update module, and a public opinion trend prediction module; the first event extraction module is used to obtain the public opinion big data of the target public opinion to be analyzed, and use the event clustering tracking module to perform event extraction and clustering processing on the public opinion big data to obtain an initial event set; wherein the initial event set contains a number of event clusters; the second event extraction module is used to obtain incremental data of the public opinion big data of the target public opinion according to a backtracking period, and the event clustering tracking module performs event extraction and clustering processing on the incremental data to obtain a number of newly added event clusters; the event set dynamic update module is used to associate each of the newly added event clusters with each event in the initial event set. The event cluster is matched and analyzed. If the match is successful, the newly added event cluster is merged into the corresponding event cluster, otherwise the newly added event cluster is added to the initial event set to obtain a dynamic event set; wherein, both the event cluster and the newly added event cluster contain event intensity values; the public opinion trend prediction module is used to input the initial event set or the dynamic event set into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the public opinion development trend report contains the public opinion development trend segment after the current moment.

[0042] An embodiment of the present invention further discloses an electronic device, comprising: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, characterized in that: the processor executes the computer program to implement the method described in the above embodiment.

[0043] An embodiment of the present invention further discloses a computer storage medium, wherein the computer storage medium stores a computer program, and is characterized in that: the computer program is executed by a processor to implement the method described in the above embodiment.

[0044] The embodiment of the present invention further discloses a computer program product. When the computer program product is run on a terminal, the terminal implements the method described in the above embodiment.

[0045] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0048] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention.

Claims

1. A method for predicting the development trend of public opinion, characterized in that The method comprises the following steps: obtaining the public opinion big data of the target public opinion to be analyzed, using the event clustering tracking module to perform event extraction and clustering processing on the public opinion big data to obtain an initial event set; wherein the initial event set contains several event clusters; obtaining the incremental data of the public opinion big data of the target public opinion according to the backtracking period, and the event clustering tracking module performs event extraction and clustering processing on the incremental data to obtain several newly added event clusters; matching and analyzing each newly added event cluster with each event cluster in the initial event set, if the match is successful, merging the newly added event cluster into the corresponding event cluster, otherwise adding the newly added event cluster into the initial event set to obtain a dynamic event set; wherein both the event cluster and the newly added event cluster contain event intensity values; inputting the dynamic event set into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein the public opinion development trend report contains the public opinion development trend segment after the current moment; Obtaining incremental data of the public opinion big data of the target public opinion according to the backtracking period, including: using the public opinion trend analysis model to perform a shallow analysis on the initial event set to obtain the public opinion type of the target public opinion, and determining a first backtracking period based on the public opinion type; calculating the number of event clusters in the initial event set, calculating a backtracking adjustment coefficient based on the number of event clusters, multiplying the backtracking adjustment coefficient by the first backtracking period to obtain a second backtracking period, and the second backtracking period is the backtracking period; obtaining the incremental data of the public opinion big data of the target public opinion according to the backtracking period.

2. A method for predicting the development trend of public opinion according to claim 1, characterized in that: The public opinion trend analysis model is trained in the following manner: collecting model training data; wherein the model training data includes public opinion industry reports, public opinion analysis cases, and public opinion analyst summary data; using the model training data to train the public opinion trend analysis model, and publishing the trained public opinion trend analysis model.

3. A method for predicting the development trend of public opinion according to claim 2, characterized in that: An event clustering tracking module is used to extract and cluster the public opinion big data to obtain an initial event set, including: the event clustering tracking module calls an AI engine to achieve: extracting event element information from the public opinion big data, the event element information includes location, person, event type, and abstractly summarizing the location, the person, and the event type to obtain event element summary information; using a clustering algorithm to cluster each of the event element summary information to achieve event classification, and obtain a number of event clusters; and, statistically analyzing each of the event element information corresponding to each of the event clusters to obtain the event intensity value; after associating each of the event clusters with the corresponding event intensity value, integrating them to obtain the initial event set.

4. A method for predicting the development trend of public opinion according to claim 3, characterized in that: The dynamic event set is input into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion, including: the dynamic event set and manually specified predicted duration data are input into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the duration of the public opinion development trend segment after the current moment in the public opinion development trend report matches the predicted duration data.

5. The method for predicting the development trend of public opinion according to claim 1, characterized in that: The method further includes: analyzing and obtaining the activity corresponding to each event cluster based on the event intensity value of each event cluster in the dynamic event set, and clearing the event clusters whose activity is lower than an activity threshold.

6. A public opinion development trend prediction system, the system is based on the method according to any one of claims 1 to 5, the system comprises a first event extraction module, a second event extraction module, an event set dynamic update module, and a public opinion trend prediction module; characterized in that: The first event extraction module is used to obtain the public opinion big data of the target public opinion to be analyzed, and uses the event clustering tracking module to extract and cluster the public opinion big data to obtain an initial event set; wherein the initial event set contains a number of event clusters; the second event extraction module is used to obtain incremental data of the public opinion big data of the target public opinion according to the backtracking period, and the event clustering tracking module extracts and clusters the incremental data to obtain a number of newly added event clusters; the event set dynamic update module is used to dynamically update each of the newly added event clusters with the initial event set. A matching analysis is performed on each event cluster in the event set. If the match is successful, the newly added event cluster is merged into the corresponding event cluster, otherwise the newly added event cluster is added to the initial event set to obtain a dynamic event set; wherein, both the event cluster and the newly added event cluster contain event intensity values; the public opinion trend prediction module is used to input the dynamic event set into the public opinion trend analysis model based on the LLM model, and the public opinion trend analysis model generates and outputs a public opinion development trend report of the target public opinion; wherein, the public opinion development trend report contains the public opinion development trend segment after the current moment.

7. An electronic device comprising: At least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 5.

8. A computer storage medium storing a computer program, characterized in that: The computer program is executed by a processor to implement the method according to any one of claims 1 to 5.

9. A computer program product comprising a computer program stored on a non-transitory computer readable medium, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images