Event importance judgment method and system based on multi-element extraction research and judgment

Abstract

The invention provides an event importance judgment method and system based on multi-element extraction study and judgment, and the method comprises the steps: obtaining a pre-collected original text data set comprising multiple pieces of text data, and carrying out the data cleaning of the text data in the original text data set, and obtaining a first text data set; clustering the text data of the first text data set to obtain a plurality of event text clusters, each event text cluster including text data for the same event; calculating scores of a plurality of event elements of each event text cluster; and calculating a total score based on the scores of the event elements of the event text cluster, and determining an importance level corresponding to the event text cluster based on the total score. According to the scheme, a large amount of text data in the cluster corresponding to each event is analyzed, factors such as the real-time propagation report condition for the corresponding event are considered, the consideration of the public opinion fermentation condition of the corresponding event is ensured, and the accuracy of judging the importance degree of the event is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of public opinion supervision, and in particular to an event importance determination method and system based on multi-element extraction research and judgment. BACKGROUND

[0002] Event importance determination is an important basis for event early warning technology. In order to cope with the increasingly complex and changing real and network public opinion environment, it is necessary to distinguish, classify and warn public opinion events through event importance determination. Through event importance determination technology, potential risk threats can be identified in advance, thereby helping organizations and individuals to take preventive measures and reduce loss risks. Event importance determination technology mainly provides users with timely and accurate information and comprehensive situation assessment through real-time monitoring and analysis of data, thereby providing strong support for rapid decision-making and emergency response. At present, there are mainly event importance determination methods based on text and keywords, event importance determination methods based on machine learning and prediction models, event importance determination methods based on spatial data analysis, event importance determination methods based on stream data analysis, and event importance determination methods based on expert systems and knowledge graphs.

[0003] The commonly used technical implementation scheme at present is the event importance determination method based on text and keywords. This method mainly analyzes text data, identifies, classifies and understands events in the text by monitoring the frequency and context information of specific keywords or phrases in the text data, and analyzes events or trends that may be related to these keywords; however, the existing technology only analyzes single text data and does not consider factors such as real-time dissemination of reports on the corresponding event, which may underestimate the public opinion fermentation of the corresponding event, resulting in errors in the importance determination of the event. SUMMARY

[0004] In view of this, the embodiments of the present application provide an event importance determination method based on multi-element extraction research and judgment to eliminate or improve one or more defects in the prior art.

[0005] One aspect of the present application provides an event importance determination method based on multi-element extraction research and judgment, the steps of which include:

[0006] Obtaining a pre-collected original text data set including multiple pieces of text data, performing data cleaning on the text data in the original text data set to obtain a first text data set;

[0007] Clustering the text data of the first text data set to obtain multiple event text clusters, each event text cluster including text data related to the same event;

[0008] Calculating the scores of multiple event elements of each event text cluster;

[0009] The total score is calculated based on the scores of multiple event elements of an event text cluster, and the importance level corresponding to the event text cluster is determined based on the total score.

[0010] The above-mentioned approach first collects a large amount of text data, removes irrelevant data through data cleaning, and then further clusters the text data, grouping text data corresponding to the same event. After clustering, each category typically still contains a large amount of text data. A comprehensive analysis is then performed on the event type, occurrence time, location, involved parties, dissemination status, and headline status of the text data within the same cluster to determine the overall score and subsequently the importance level, providing a reference for data supervision. This approach can analyze a large amount of text data within each event's corresponding cluster, taking into account factors such as real-time dissemination and reporting of the event, ensuring consideration of public opinion development and improving the accuracy of event importance determination.

[0011] In some embodiments of the present invention, in the step of cleaning the text data in the original text data set to obtain the first text data set, regular expressions are used for data cleaning.

[0012] In some embodiments of the present invention, in the step of clustering the text data of the first text data set to obtain multiple event text clusters, a pre-trained first neural network model is used to cluster the text data of the first text data set.

[0013] In some embodiments of the present invention, after obtaining a pre-collected original text data set including multiple text data, performing data cleaning on the text data in the original text data set to obtain a first text data set, the method further includes the step of using a pre-trained second neural network model to classify the text data in the first text data set into event text and non-event text, deleting the non-event text in the first text data set, and obtaining an updated first text data set.

[0014] In some embodiments of the present invention, the event elements include event type, time of occurrence, location of occurrence, parties involved, dissemination status, and headline status. The step of calculating the score of multiple event elements for each event text cluster includes:

[0015] The text data is classified based on a pre-trained third neural network model to obtain the category of the text data, and the score of the event text cluster for the event type is determined based on the category of the text data.

[0016] Based on the text extraction algorithm, the occurrence time, location, and parties involved in the text data are extracted. Based on the extracted occurrence time, location, and parties involved, the score of the text cluster of the event is determined for the occurrence time, location, and parties involved.

[0017] The score of the event text cluster for its propagation status is calculated based on the propagation data of the text data in the event text cluster.

[0018] The score for the event text cluster is determined based on the number of headlines in the text data of that event text cluster.

[0019] In some embodiments of the present invention, in the step of calculating the total score based on the scores of multiple event elements of an event text cluster, a weighted score of each event element is calculated using a weighted calculation method based on the weight value corresponding to each event element, and the total score is calculated based on the weighted scores of each event element.

[0020] In some embodiments of the present invention, in the step of calculating the total score based on the weighted scores of each event element, the total score is obtained by adding the weighted scores of each event element or by adding the weighted scores of event type, occurrence time, occurrence location, involved parties and dissemination status to obtain an initial score, and the total score is calculated based on the initial score and the weighted score of the headline situation.

[0021] In some embodiments of the present invention, the database records data on historical hot events, including the event type, occurrence time, location, and involved parties. The step of calculating the total score based on multiple event elements of the event text cluster further includes:

[0022] The score for each historical hot topic event is determined based on its event type, time, location, and involved parties.

[0023] The event type, time of occurrence, location of occurrence, and the score value corresponding to the parties involved in each historical hot event in the database are used to construct a matching vector;

[0024] Construct a matching vector from the event text clusters, including the event type, time of occurrence, location of occurrence, and the rating values ​​corresponding to the parties involved.

[0025] Calculate the distance between the matching vector and each unmatched vector. If the distance between the matching vector and any unmatched vector is less than a preset distance threshold, then the weight combination corresponding to the unmatched vector is applied to the step of calculating the total score of multiple event elements based on the event text cluster.

[0026] In some embodiments of the present invention, if the distance between the matching vector and any unmatched vector is not less than a preset distance threshold, then the preset weight combination is applied to the step of calculating the total score of multiple event elements based on the event text cluster.

[0027] In some embodiments of the present invention, the weight combination includes a weight value corresponding to each event element. The weight combination corresponding to the vector to be matched is applied to the step of calculating the total score of multiple event elements based on the event text cluster. The weight values ​​corresponding to each event element in the weight combination are used to calculate the weighted score of each event element in a weighted calculation method.

[0028] A second aspect of the present invention also provides an event importance determination system based on multi-factor extraction and analysis. The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method described above.

[0029] A third aspect of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned event importance determination method based on multi-factor extraction and analysis.

[0030] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the text, or may be learned by practice of the invention. The objects and other advantages of the invention will become apparent from the description and the accompanying drawings.

[0031] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0032] The accompanying drawings, which are provided to further illustrate the invention and form part of this application, are not intended to limit the scope of the invention.

[0033] Figure 1 This is a schematic diagram of the first embodiment of the event importance determination method based on multi-factor extraction and analysis of the present invention;

[0034] Figure 2 This is a schematic diagram of the second implementation of the event importance determination method based on multi-factor extraction and analysis of the present invention;

[0035] Figure 3This is a schematic diagram of the third implementation of the event importance determination method based on multi-factor extraction and analysis of the present invention;

[0036] Figure 4 This is a schematic diagram of the overall architecture of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0038] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0039] Highway transportation, as one of the most important modes of road transport in China, plays a supporting role in socio-economic development. However, safety issues related to key vehicles such as medium and large passenger buses and hazardous materials vehicles are prominent. Therefore, in response to the need for intelligent traffic monitoring, it is necessary to establish a system for monitoring and predicting the road traffic of key vehicles to prevent major traffic accidents involving these vehicles and avoid significant casualties.

[0040] like Figure 1 and 4 As shown, this invention proposes a method for determining the importance of events based on multi-factor extraction and analysis. The steps of this method include:

[0041] Step S100: Obtain the pre-collected original text data set including multiple text data, and perform data cleaning on the text data in the original text data set to obtain the first text data set;

[0042] In the specific implementation process, the pre-collected raw text data set includes multiple text data, which is a collection of text data obtained from the Internet over a historical period of time.

[0043] Step S200: Cluster the text data of the first text data set to obtain multiple event text clusters, each of which includes text data related to the same event.

[0044] In the specific implementation process, in the step of clustering the text data of the first text data set using a pre-trained first neural network model, commonly used neural network models include autoencoders, convolutional neural networks (CNNs), recurrent neural networks (RNNs) and their variants (such as LSTM and GRU), and clustering algorithms can be K-means, hierarchical clustering, or DBSCAN, etc.

[0045] Step S300: Calculate the scores of multiple event elements for each event text cluster;

[0046] Step S400: Calculate the total score based on the scores of multiple event elements of the event text cluster, and determine the importance level corresponding to the event text cluster based on the total score.

[0047] like Figure 2 and 3 As shown, in the specific implementation process, the step of determining the importance level corresponding to the event text cluster based on the total score includes step S460, where each importance level corresponds to a scoring interval, and the importance level of the event text cluster is determined based on the scoring interval corresponding to the total score of the event text cluster.

[0048] The above-mentioned approach first collects a large amount of text data, removes irrelevant data through data cleaning, and then further clusters the text data, grouping text data corresponding to the same event. After clustering, each category typically still contains a large amount of text data. A comprehensive analysis is then performed on the event type, occurrence time, location, involved parties, dissemination status, and headline status of the text data within the same cluster to determine the overall score and subsequently the importance level, providing a reference for data supervision. This approach can analyze a large amount of text data within each event's corresponding cluster, taking into account factors such as real-time dissemination and reporting of the event, ensuring consideration of public opinion development and improving the accuracy of event importance determination.

[0049] In some embodiments of the present invention, in the step of cleaning the text data in the original text data set to obtain the first text data set, regular expressions are used for data cleaning.

[0050] In practical implementation, during the data cleaning process using regular expressions, regular expressions (Regex) are a powerful text processing tool used to match character combinations in a string that conform to a specific pattern. Through predefined character sets, character classes, quantifiers, and other rules, they can accurately find, replace, or delete specific content in text. During text cleaning, Regex can help remove useless characters, correct formatting errors, and extract key information, greatly improving the efficiency and accuracy of data processing.

[0051] To improve computational efficiency and event detection accuracy using the above approach, it is necessary to prioritize cleaning the original data, removing emojis, special characters, hyperlinks, and other content that are not long enough to present complete semantic information.

[0052] In some embodiments of the present invention, in the step of clustering the text data of the first text data set to obtain multiple event text clusters, a pre-trained first neural network model is used to cluster the text data of the first text data set.

[0053] Using the above scheme, since there are usually multiple reports and posts about the same event in cyberspace, and the event information contained in different reports or posts usually has certain differences, event fusion can effectively fill in the missing event information in data from different sources. At the same time, merging multiple event data into an event cluster for subsequent calculation can effectively improve the calculation efficiency.

[0054] Text clustering groups texts based on similarity, classifying similar texts into the same category, thus achieving effective aggregation of event description texts. Clustering algorithms are unsupervised algorithms, requiring no large amounts of labeled data, and text clustering can discover potential relationships and commonalities between different event texts, making it well-suited for event fusion using text data. However, this approach has relatively high computational requirements; as the amount of text increases, computational consumption increases exponentially. Therefore, the choice of approach should be based on the specific data requirements. Besides text clustering, event fusion can also be performed by calculating the similarity of core features such as the event's time, location, and participating entities.

[0055] In some embodiments of the present invention, after obtaining a pre-collected original text data set including multiple text data, performing data cleaning on the text data in the original text data set to obtain a first text data set, the method further includes the step of using a pre-trained second neural network model to classify the text data in the first text data set into event text and non-event text, deleting the non-event text in the first text data set, and obtaining an updated first text data set.

[0056] In the specific implementation process, event text is filtered through event detection. Event detection is a computational operation that uses natural language processing (NLP) technologies to analyze data from news media and social media platforms to achieve automatic event identification. This solution targets text-based data. The content of text data from online platforms, especially social media posts, is highly complex. Besides event descriptions, it also includes a large amount of irrelevant text such as emotional expressions, advertisements, discussions of cultural and creative content, and game guides. To effectively identify business-related events from massive amounts of data, this solution constructs a targeted event detection model based on generative artificial intelligence (GI) to achieve accurate detection of business-related events. In addition to building a second neural network model based on generative AI, a second neural network model can also be built based on machine learning, such as random forests, XGBoost, and multilayer neural networks. Building a second neural network model based on machine learning requires more labeled data, while building an event detection model based on generative AI requires less labeled data but has higher computational requirements and is less efficient than machine learning models. The choice can be made based on the actual business application needs.

[0057] In some embodiments of the present invention, the event elements include event type, time of occurrence, location of occurrence, parties involved, dissemination status, and headline status. The step of calculating the score of multiple event elements for each event text cluster includes:

[0058] The text data is classified based on a pre-trained third neural network model to obtain the category of the text data, and the score of the event text cluster for the event type is determined based on the category of the text data.

[0059] In practice, each category of text data corresponds to a preset score.

[0060] In the specific implementation process, the average event type score of the text data in the event text cluster is used as the score of the event text cluster for the event type.

[0061] Based on the text extraction algorithm, the occurrence time, location, and parties involved in the text data are extracted. Based on the extracted occurrence time, location, and parties involved, the score of the text cluster of the event is determined for the occurrence time, location, and parties involved.

[0062] In the specific implementation process, Named Entity Recognition (NER) technology can be used to extract relevant elements to obtain the occurrence time, location and involved parties of the text data. This scheme has high computational efficiency, but due to the lack of deep understanding of the text content, the extraction accuracy is relatively low, and it is easy to extract irrelevant elements. For example, multiple locations may appear in the event description text, and it is difficult to determine which location is the actual location of the event based solely on named entity recognition.

[0063] In the specific implementation process, a dedicated event core element extraction model can also be built based on open source large models such as Llama to extract relevant elements. After testing, the basic event features extracted by this solution have high accuracy, but it has high computing power requirements. When computing power is limited, it is only suitable for extracting basic event features when the amount of data is relatively small.

[0064] In the specific implementation process, in addition to the features mentioned above, the basic characteristics of an event also include features based on data collection or statistical acquisition, such as the platform on which the event was first reported, the reporting time, the average length of the posts, and the data modality.

[0065] In the specific implementation process, the average score of the occurrence time, occurrence location, and involved parties in the text data of the event text cluster is used as the score of the event text cluster for the occurrence time, occurrence location, and involved parties.

[0066] The score of the event text cluster for its propagation status is calculated based on the propagation data of the text data in the event text cluster.

[0067] In the specific implementation process, a weighted calculation is performed based on the number of reposts, comments, and likes of the text data in the event text cluster to obtain a score for the dissemination of the event text cluster.

[0068] The score for the event text cluster is determined based on the number of headlines in the text data of that event text cluster.

[0069] In practice, different numbers of headlines correspond to different scores. The score is determined based on the number of headlines in the text data of the event text cluster. Specifically, when there are more than 3 headline reports, the score is 4; when there are 1-2 headline reports, the score is 2; and when there are no headline reports, the score is 1.

[0070] like Figure 2 and 3 As shown, in some embodiments of the present invention, the step of calculating the total score based on the scores of multiple event elements of an event text cluster includes step S450, calculating the weighted score of each event element based on the weight value corresponding to each event element using a weighted calculation method, and calculating the total score based on the weighted score of each event element.

[0071] In some embodiments of the present invention, in the step of calculating the total score based on the weighted scores of each event element, the total score is obtained by adding the weighted scores of each event element or by adding the weighted scores of event type, occurrence time, occurrence location, involved parties and dissemination status to obtain an initial score, and the total score is calculated based on the initial score and the weighted score of the headline situation.

[0072] In practice, if the weighted scores of each event element are summed to obtain the total score, the total score is calculated based on the following formula:

[0073]

[0074] Where R represents the total score, i represents any event element, and x i w represents the score of any event element. i represents the weight value of any event element, and n represents the total number of event elements.

[0075] In the specific implementation process, if an initial score is obtained by adding weighted scores based on event type, time of occurrence, location of occurrence, involved parties, and dissemination status, and the total score is calculated based on the weighted score of the initial score and the Toutiao (ByteDance's news aggregator) data, then the total score is calculated using the following formula:

[0076] R=(w1·P+w2·L+w3·C+w4·S+w5·A)·w6·B;

[0077] Among them, w1, w2, w3, w4, w5 and w6 represent the weight values ​​corresponding to the event type, occurrence time, occurrence location, involved parties, dissemination status and headline status, respectively, and P, L, C, S, A and B represent the scores corresponding to the event type, occurrence time, occurrence location, involved parties, dissemination status and headline status, respectively.

[0078] like Figure 2 and 3 As shown, in some embodiments of the present invention, the database records data on historical hot events. This data includes the event type, occurrence time, location, and involved parties. The step of calculating the total score based on multiple event elements of the event text cluster further includes:

[0079] Step S410: Determine the score corresponding to the event type, time of occurrence, location of occurrence, and parties involved in each historical hot event based on the event type, time of occurrence, location of occurrence, and parties involved.

[0080] In the specific implementation process, in the step of determining the score corresponding to the event type, time, location and involved parties of each historical hot event based on the event type, time, location and involved parties of the historical hot event, the same method as step S300 is used to determine the score.

[0081] Step S420: Construct a matching vector from the event type, occurrence time, occurrence location, and the score value corresponding to the involved parties for each historical hot event in the database;

[0082] Step S430: Construct a matching vector from the event text clusters corresponding to the event type, occurrence time, occurrence location, and the score of the involved parties;

[0083] Step S440: Calculate the distance between the matching vector and each vector to be matched;

[0084] In the specific implementation process, in the step of calculating the distance between the matching vector and each vector to be matched, the distance value is calculated using methods such as cosine distance or Euclidean distance.

[0085] Step S441: If the distance between the matching vector and any unmatched vector is less than a preset distance threshold, then the weight combination corresponding to the unmatched vector is applied to the step of calculating the total score of multiple event elements based on the event text cluster.

[0086] In some embodiments of the present invention, in step S442, if the distance between the matching vector and any unmatched vector is not less than a preset distance threshold, then the preset weight combination is applied to the step of calculating the total score of multiple event elements based on the event text cluster.

[0087] Using the above scheme, since historical hot events are relatively popular times in the past, and subsequent events are often related to the earlier historical hot events, this scheme determines the relevance between the current event and each historical hot event by calculating the distance between the matching vector and each unmatched vector. If the relevance is high, the weight combination set for the historical hot event is applied to the calculation of the total score, and the final total score reflects the relevance between the event and the historical hot events.

[0088] In the step of applying a preset weight combination to the scoring of multiple event elements based on event text clusters to calculate the total score, the preset weight combination is a pre-defined weight combination used for situations unrelated to historical hot events.

[0089] In some embodiments of the present invention, the weight combination includes a weight value corresponding to each event element. The weight combination corresponding to the vector to be matched is applied to the step of calculating the total score of multiple event elements based on the event text cluster. The weight values ​​corresponding to each event element in the weight combination are used to calculate the weighted score of each event element in a weighted calculation method.

[0090] This approach, employing the aforementioned scheme, combines basic event characteristics, initial dissemination characteristics, event type characteristics, and headline reporting characteristics. Through real-time monitoring of data from news media and social media platforms, it achieves scientific assessment and tiered early warning of event risk. Research shows that this method has high accuracy and effectiveness in event detection and risk assessment. Although there is still significant room for optimization in core event recall, the model as a whole demonstrates strong practicality and technical feasibility. Future research can further optimize the algorithms for event detection, feature extraction, and event type identification, and dynamically adjust the weight parameters of different event characteristics according to actual business needs to further improve the model's accuracy and application breadth. This research scheme not only provides a new technical path for event early warning but also offers strong support for risk prevention and control in the fields of social security and public management, possessing extremely high social application value and broad application prospects.

[0091] This invention also provides an event importance determination system based on multi-factor extraction and analysis. The system includes a computer device, which includes a processor and a memory. The memory stores computer instructions, and the processor executes the computer instructions stored in the memory. When the computer instructions are executed by the processor, the system implements the steps of the method described above.

[0092] This invention also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the steps of the aforementioned event importance determination method based on multi-factor extraction and analysis. The computer-readable storage medium can be a tangible storage medium, such as random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, register, floppy disk, hard disk, removable storage disk, CD-ROM, or any other form of storage medium known in the art.

[0093] Those skilled in the art will understand that the exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention. When implemented in hardware, it can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the desired tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave.

[0094] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0095] In this invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or in place of features of other embodiments.

[0096] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An event importance determination method based on multi-element extraction research and judgment, characterized in that, The method comprises the following steps: An original text data set including a plurality of text data is acquired, and data cleaning is performed on the text data in the original text data set to obtain a first text data set; The text data in the first text data set is clustered to obtain a plurality of event text clusters, and each event text cluster includes text data of the same event; The scores of a plurality of event elements of each event text cluster are calculated; The total score is calculated based on the scores of the plurality of event elements of the event text cluster, and the importance level corresponding to the event text cluster is determined based on the total score.

2. The event importance determination method based on multi-element extraction research and judgment according to claim 1, characterized in that, In the step of performing data cleaning on the text data in the original text data set to obtain the first text data set, the data cleaning is performed in the form of a regular expression. 3.The event importance determination method based on multi-element extraction research and judgment according to claim 1, wherein, In the step of clustering the text data in the first text data set to obtain the plurality of event text clusters, the first pre-trained neural network model is used to cluster the text data in the first text data set.

4. The event importance determination method based on multi-element extraction research and judgment according to claim 1, characterized in that, After the step of acquiring the original text data set including a plurality of text data, performing data cleaning on the text data in the original text data set to obtain the first text data set, the step of using the second pre-trained neural network model to classify the text data in the first text data set into event text and non-event text, deleting the non-event text in the first text data set, and obtaining an updated first text data set is further included.

5. The method of claim 1, wherein the method further comprises: The event elements include event type, occurrence time, occurrence place, involved subject, propagation condition and headline condition, and the step of calculating the scores of the plurality of event elements of each event text cluster comprises: The text data is classified based on the third pre-trained neural network model to obtain the category of the text data, and the score of the event type of the event text cluster is determined based on the category of the text data; The occurrence time, occurrence place and involved subject of the text data are extracted based on the text extraction algorithm, and the scores of the occurrence time, occurrence place and involved subject of the event text cluster are determined based on the extracted occurrence time, occurrence place and involved subject; The propagation data of the text data of the event text cluster are calculated to obtain the score of the propagation condition of the event text cluster; The score of the headline condition of the event text cluster is determined based on the number of headlines of the text data of the event text cluster.

6. The method of claim 1, wherein the method further comprises: In the step of calculating the total score based on the scores of the plurality of event elements of the event text cluster, the weighted scores of the plurality of event elements are calculated based on the weight values corresponding to the plurality of event elements, and the total score is calculated based on the weighted scores of the plurality of event elements.

7. The event importance determination method based on multi-element extraction research and judgment according to claim 6, characterized in that, In the step of calculating the total score based on the weighted scores of the plurality of event elements, the total score is obtained by adding the weighted scores of the plurality of event elements, or the initial score is obtained by adding the weighted scores of the event type, occurrence time, occurrence place, involved subject and propagation condition, and the total score is calculated based on the initial score and the weighted score of the headline condition.

8. The event importance determination method based on multi-element extraction analysis according to any one of claims 1 to 7, characterized by, The database records data of historical hot events, the data of the historical hot events including event types, occurrence times, occurrence places and involved subjects of historical hot data, and the step of calculating the total score based on the scores of the multiple event elements of the event text cluster further includes: determining the scores corresponding to the event types, occurrence times, occurrence places and involved subjects of each historical hot event based on the event types, occurrence times, occurrence places and involved subjects of the historical hot events; constructing the values of the scores corresponding to the event types, occurrence times, occurrence places and involved subjects of each historical hot event in the database into a to-be-matched vector; constructing the values of the scores corresponding to the event types, occurrence times, occurrence places and involved subjects in the event text cluster into a matched vector; calculating the distances between the matched vector and each to-be-matched vector, and if the distance between the matched vector and any to-be-matched vector is less than a preset distance threshold, then applying the weight combination corresponding to the to-be-matched vector to the step of calculating the total score based on the scores of the multiple event elements of the event text cluster.

9. The event importance determination method based on multi-element extraction research and judgment according to claim 8, characterized in that, If the distances between the matched vector and any to-be-matched vector are all not less than the preset distance threshold, then applying a preset weight combination to the step of calculating the total score based on the scores of the multiple event elements of the event text cluster.

10. An event importance determination system based on multi-element extraction research and judgment, characterized in that, The system includes a computer device including a processor and a memory, the memory storing computer instructions, and the processor is configured to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the system implements the steps as claimed in any one of claims 1-9.

Images