Risk event analysis method and system based on data analysis

Abstract

This invention discloses a risk event analysis method and system based on data analysis, belonging to the field of data mining technology. The technical problem it addresses is how to overcome the inefficiency of manually screening risk event information from large amounts of data. The method includes: acquiring historical risk data sets and current event text; establishing an index for historical risk data; extracting high-frequency keywords from the current event text; constructing a risk analysis hot word set based on the core high-frequency keywords; obtaining historical correlation data related to the current event from the historical risk data index; determining the risk hazard based on the current event text and historical correlation data based on the risk analysis hot word set; quantifying and scoring the risk level based on the risk hazard determination results; and issuing risk warnings based on the quantified risk level scores.

Description

Technical Field

[0001] This invention relates to the field of data mining technology, specifically to a risk event analysis method and system based on data analysis. Background Technology

[0002] With the rapid development of modern society, the interests of individuals, groups, and other parties are intertwined and complex, and incidents such as conflicts and disputes are emerging one after another. How to efficiently mine high-risk data from a large amount of data and provide real-time early warnings is of great significance for preventing the escalation of risks, maintaining social stability, and assisting judicial, community and other departments to improve their work efficiency.

[0003] Overcoming the inefficiency of manually screening risk event information from large amounts of data is a technical problem that needs to be solved. Summary of the Invention

[0004] The technical objective of this invention is to address the above-mentioned shortcomings by providing a risk event analysis method and system based on data analysis, thereby solving the technical problem of low efficiency in manually screening risk event information from large amounts of data.

[0005] In a first aspect, the present invention provides a risk event analysis method based on data analysis, comprising the following steps: Risk Analysis: Acquire historical risk data sets and current event texts, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event texts, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data consists of text records of various risk events that have occurred, and current event texts are the target event texts to be analyzed. Fusion Analysis: Based on the index of historical risk data, historical related data associated with the current event is obtained from the historical risk data. Based on the risk analysis hot word set, the current event text and historical related data are used to determine the risk hazards. Based on the risk hazard determination results, the risk level is quantitatively scored, and the risk warning is issued based on the risk level quantitative scoring results.

[0006] As a preferred approach, risk analysis includes the following steps: Index building: Import historical risk data into the full-text search engine and build a structured index using the indexing function of the full-text search engine; Data preprocessing: The current event text is cleaned to remove noise data, including stop words and punctuation marks, to obtain the preprocessed text; Text segmentation: Based on a relevant domain dictionary, the preprocessed text is segmented using a segmentation algorithm to break down continuous text content into independent lexical units, resulting in segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the field to which the risk event belongs. Keyword extraction: Core high-frequency keywords that are indicative of risk events were screened from the word segmentation results using the TF-IDF statistical method; Hot keyword collection construction: Integrate and classify core high-frequency keywords to form a structured hot keyword collection for risk analysis.

[0007] As a preferred method, when selecting core high-frequency keywords that are indicative of risk events from the word segmentation results using the TF-IDF statistical method, the word frequency and inverse document frequency of each word unit after word segmentation are calculated, and the word frequency and inverse document frequency of each word unit are combined for calculation. Based on the calculation results, core high-frequency keywords that are highly indicative of risk events are selected. Among them, word frequency represents the frequency of a word unit in the current text, and inverse document frequency represents the inverse frequency of a word unit in the historical risk dataset.

[0008] As a preferred approach, fusion analysis includes the following operations: Historical correlation data mining: Using the information of personnel related to the current event as the correlation key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical correlation data, and the cumulative frequency of occurrence of historical correlation data is counted; Risk and hazard assessment: Based on the risk hot word set, the current event text and historical related data are compared and analyzed. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hazards in the current event. Risk data weighting percentage calculation: The percentage of total risk data to sample data is calculated to obtain the risk data weighting percentage. Risk data includes current event text identified as risk data and historical related data identified as risk data. Sample data includes current event text and historical related data. Comprehensive quantitative scoring: Based on the proportion of risk data, scores are calculated according to preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the corresponding score is added; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the corresponding score is added; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the corresponding score is added. Classification and early warning output: The risk level is determined according to the cumulative score and a classification and early warning is executed. If the final score exceeds the predetermined value, the early warning information of historical related data is output simultaneously while the current event risk warning is output.

[0009] Secondly, the present invention provides a risk event analysis system based on data analysis, comprising a risk analysis module and a fusion analysis module; The risk analysis module is used to perform the following: acquire historical risk data sets and current event texts, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event texts, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data are text records of various risk events that have occurred, and current event texts are target event texts to be analyzed. The fusion analysis module is used to perform the following: based on the index of historical risk data, it retrieves historical related data associated with the current event from the historical risk data; based on the risk analysis hot word set, it determines the risk hazards of the current event text and historical related data; based on the risk hazard determination results, it performs risk level quantification scoring; and based on the risk level quantification scoring results, it provides risk warnings.

[0010] As a preferred option, the risk analysis module is used to perform the following operations: Index building: Import historical risk data into the full-text search engine and build a structured index using the indexing function of the full-text search engine; Data preprocessing: The current event text is cleaned to remove noise data, including stop words and punctuation marks, to obtain the preprocessed text; Text segmentation: Based on a relevant domain dictionary, the preprocessed text is segmented using a segmentation algorithm to break down continuous text content into independent lexical units, resulting in segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the field to which the risk event belongs. Keyword extraction: Core high-frequency keywords that are indicative of risk events were screened from the word segmentation results using the TF-IDF statistical method; Hot keyword collection construction: Integrate and classify core high-frequency keywords to form a structured hot keyword collection for risk analysis.

[0011] As a preferred method, when selecting core high-frequency keywords that are indicative of risk events from the word segmentation results using the TF-IDF statistical method, the word frequency and inverse document frequency of each word unit after word segmentation are calculated, and the word frequency and inverse document frequency of each word unit are combined for calculation. Based on the calculation results, core high-frequency keywords that are highly indicative of risk events are selected. Among them, word frequency represents the frequency of a word unit in the current text, and inverse document frequency represents the inverse frequency of a word unit in the historical risk dataset.

[0012] Preferably, the fusion analysis module is used to perform the following operations: Historical correlation data mining: Using the information of personnel related to the current event as the correlation key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical correlation data, and the cumulative frequency of occurrence of historical correlation data is counted; Risk and hazard assessment: Based on the risk hot word set, the current event text and historical related data are compared and analyzed. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hazards in the current event. Risk data weighting percentage calculation: The percentage of total risk data to sample data is calculated to obtain the risk data weighting percentage. Risk data includes current event text identified as risk data and historical related data identified as risk data. Sample data includes current event text and historical related data. Comprehensive quantitative scoring: Based on the proportion of risk data, scores are calculated according to preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the corresponding score is added; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the corresponding score is added; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the corresponding score is added. Classification and early warning output: The risk level is determined according to the cumulative score and a classification and early warning is executed. If the final score exceeds the predetermined value, the early warning information of historical related data is output simultaneously while the current event risk warning is output.

[0013] The risk event analysis method and system based on data analysis of the present invention have the following advantages: longitudinal comparison with related historical data for risk analysis, mining potential risks from the data itself and related data including frequency information, providing early warning of high-risk and high-frequency event data analysis, reducing manual intervention, and improving the problem of low efficiency of manual screening. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] The invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 This is a flowchart of a risk event analysis method based on data analysis, as shown in Example 1. Detailed Implementation

[0017] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments are not intended to limit the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0018] This invention provides a risk event analysis method and system based on data analysis, which addresses the technical problem of low efficiency in manually screening risk event information from large amounts of data. Example

[0019] This invention provides a risk event analysis method based on data analysis, which includes two steps: risk analysis and fusion analysis.

[0020] The terms used in this embodiment are defined as follows: Risk events: Events that, in themselves or in their related data, pose a potential risk that could materialize and lead to actual losses or worsen the situation. Conflicts and disputes: Conflicts and disputes are disputes and conflicts between different subjects in social life due to conflicts of interest and disputes over rights. The subjects are often diverse and have potential social risks, such as neighborhood disputes, labor disputes and debt disputes. Risk keywords: Keywords that frequently appear in high-risk data and have certain potential hidden dangers, which are likely to exacerbate actual conflicts, such as threats and harms; High-frequency data: In historical data of the database, correlation analysis is performed based on the subject or content of the current event. Data with a certain amount of correlation is considered high-frequency data. Sample weighting percentage: The percentage of the target result in the total sample.

[0021] Step S100 Risk Analysis: Obtain historical risk data set and current event text, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event text, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data is the text record of various risk events that have occurred, and current event text is the target event text to be analyzed.

[0022] As a specific implementation of risk analysis, this step includes the following operations: (1) Index building: Import historical risk data into the full-text search engine and build a structured index through the indexing function of the full-text search engine; (2) Data preprocessing: Data cleaning is performed on the current event text to remove noise data, including stop words and punctuation marks, and the preprocessed text is obtained; (3) Text segmentation: Based on the relevant domain dictionary, the preprocessed text is segmented by the segmentation algorithm to split the continuous text content into independent lexical units and obtain the segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the domain to which the risk event belongs. (4) Keyword extraction: Core high-frequency keywords that are indicative of risk events are selected from the word segmentation results using the TF-IDF statistical method; (5) Construction of hot word set: Integrate and classify core high-frequency keywords to form a structured risk analysis hot word set.

[0023] In this embodiment, when selecting core high-frequency keywords that are indicative of risk events from the word segmentation results using the TF-IDF statistical method, the term frequency and inverse document frequency (IVF) of each word unit after segmentation are calculated. The term frequency and IVF of each word unit are then combined for calculation, and core high-frequency keywords with strong indicative power for risk events are selected based on the calculation results. Here, term frequency represents the frequency of a word unit appearing in the current text, and IVF represents the reverse frequency of a word unit appearing in the historical risk data set.

[0024] In this embodiment, the risk analysis involves indexing historical risk data using a full-text search engine, cleaning the event content itself to remove stop words, punctuation, and other noise, segmenting the text content using relevant domain dictionaries, and extracting core high-frequency keywords using the TF-IDF statistical method through term frequency-inverse document frequency to construct a set of hot words for risk analysis, which serves as the basis for risk analysis.

[0025] Step S200 Fusion Analysis: Based on the index of historical risk data, retrieve historical related data associated with the current event from the historical risk data, determine the risk hazards of the current event text and historical related data based on the risk analysis hot word set, quantify and score the risk level based on the risk hazard determination results, and issue risk warnings based on the risk level quantification and scoring results.

[0026] As a specific implementation of the fusion analysis, this step includes the following operations: (1) Historical association data mining: Using the information of personnel related to the current event as the association key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical association data, and the cumulative frequency of occurrence of historical association data is counted; (2) Risk and hidden danger determination: Based on the risk hot word set, the current event text and historical related data are compared and analyzed respectively. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hidden dangers in the current event. (3) Calculation of risk data proportion: The total amount of risk data accounts for the percentage of sample data to obtain the risk data proportion. Risk data includes current event text that is identified as risk data and historical related data that is identified as risk data. Sample data includes current event text and historical related data. (4) Comprehensive quantitative scoring: Based on the proportion of risk data, the score is calculated according to the preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the score is increased accordingly; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the score is increased accordingly; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the score is increased accordingly. (5) Classification warning output: The risk level is determined according to the cumulative score and classification warning is executed. If the final score exceeds the predetermined score, the warning information of historical related data is output simultaneously while the current event risk warning is output.

[0027] This embodiment utilizes Hadoop MapReduce distributed computing for fusion analysis. Based on information about individuals related to the current event, such as contact information, it mines historical relevance data from massive datasets and counts the frequency of historical occurrences. Risk analysis is performed separately for the current event itself and historical related data to determine potential risks and calculate the percentage of risky data in the total sample data. This is combined with the frequency of historical related data for a comprehensive analysis. If the current data itself is determined to be risky through risk analysis, the score is increased by 3. If the frequency of historical related data exceeds a certain threshold, the score is increased by 2. If the proportion of risky data in historical related data exceeds a certain threshold, the score is increased by 1. Based on the final score and the determination results, a classification and warning are issued. If the score exceeds 3 points, a warning is simultaneously issued for its historical related data.

[0028] In this embodiment, after step S100 is completed, it provides a basis for step S200 to determine whether the risk analysis belongs to a risk event. After step S200 is completed, it determines whether it is determined to be a risk event, and the final result of the warning after determining it to be a risk event includes historical related data.

[0029] The method disclosed in this embodiment, in specific business scenarios such as the investigation of disputes, utilizes collected historical data to generate risk classification criteria through word segmentation and word frequency statistics. Based on the original data to be analyzed, historical correlation data is mined through big data mining, and risk analysis is performed separately. Finally, based on the historical frequency of data occurrence and risk classification, a comprehensive analysis of risk data is achieved and early warning is provided. Example

[0030] This invention provides a risk event analysis system based on data analysis, comprising a risk analysis module and a fusion analysis module.

[0031] The risk analysis module is used to perform the following: acquire historical risk data sets and current event texts, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event texts, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data are text records of various risk events that have occurred, and current event texts are target event texts to be analyzed.

[0032] As a specific implementation of the risk analysis module, this module is used to perform the following operations: (1) Index building: Import historical risk data into the full-text search engine and build a structured index through the indexing function of the full-text search engine; (2) Data preprocessing: Data cleaning is performed on the current event text to remove noise data, including stop words and punctuation marks, and the preprocessed text is obtained; (3) Text segmentation: Based on the relevant domain dictionary, the preprocessed text is segmented by the segmentation algorithm to split the continuous text content into independent lexical units and obtain the segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the domain to which the risk event belongs. (4) Keyword extraction: Core high-frequency keywords that are indicative of risk events are selected from the word segmentation results using the TF-IDF statistical method; (5) Construction of hot word set: Integrate and classify core high-frequency keywords to form a structured risk analysis hot word set.

[0033] In this embodiment, when selecting core high-frequency keywords that are indicative of risk events from the word segmentation results using the TF-IDF statistical method, the term frequency and inverse document frequency (IVF) of each word unit after segmentation are calculated. The term frequency and IVF of each word unit are then combined for calculation, and core high-frequency keywords with strong indicative power for risk events are selected based on the calculation results. Here, term frequency represents the frequency of a word unit appearing in the current text, and IVF represents the reverse frequency of a word unit appearing in the historical risk data set.

[0034] In this embodiment, the risk analysis module is used to index historical risk data through a full-text search engine, clean the event content itself to remove stop words, punctuation and other noise data, segment the text content with relevant domain dictionaries, and use the TF-IDF statistical method to extract core high-frequency keywords through word frequency-inverse document frequency to construct a risk analysis hot word set as the basis for risk analysis.

[0035] The fusion analysis module is used to perform the following: based on the index of historical risk data, it retrieves historical related data associated with the current event from the historical risk data; based on the risk analysis hot word set, it determines the risk hazards of the current event text and historical related data; based on the risk hazard determination results, it performs risk level quantification scoring; and based on the risk level quantification scoring results, it provides risk warnings.

[0036] As a specific implementation of the fusion analysis module, this module is used to perform the following operations: (1) Historical association data mining: Using the information of personnel related to the current event as the association key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical association data, and the cumulative frequency of occurrence of historical association data is counted; (2) Risk and hidden danger determination: Based on the risk hot word set, the current event text and historical related data are compared and analyzed respectively. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hidden dangers in the current event. (3) Calculation of risk data proportion: The total amount of risk data accounts for the percentage of sample data to obtain the risk data proportion. Risk data includes current event text that is identified as risk data and historical related data that is identified as risk data. Sample data includes current event text and historical related data. (4) Comprehensive quantitative scoring: Based on the proportion of risk data, the score is calculated according to the preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the score is increased accordingly; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the score is increased accordingly; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the score is increased accordingly. (5) Classification warning output: The risk level is determined according to the cumulative score and classification warning is executed. If the final score exceeds the predetermined score, the warning information of historical related data is output simultaneously while the current event risk warning is output.

[0037] This embodiment's fusion analysis module utilizes Hadoop MapReduce distributed computing to mine historical relevance data from massive datasets based on information about individuals involved in the current event, such as contact information, and to statistically analyze the historical frequency of occurrence. Risk analysis is performed on both the current event itself and historical related data to determine potential risks and calculate the percentage of risky data in the total sample data. This is combined with the frequency of historical related data for a comprehensive analysis. If the current data itself is determined to be risky through risk analysis, the score is increased by 3. If the frequency of historical related data exceeds a certain threshold, the score is increased by 2. If the proportion of risky data in historical related data exceeds a certain threshold, the score is increased by 1. Based on the final score and the determination results, a classification and warning are issued. If the score exceeds 3 points, a warning is simultaneously issued for its historical related data.

[0038] The system in this embodiment can perform risk event analysis by executing the method disclosed in Embodiment 1.

[0039] The above provides a detailed description of the risk event analysis method and system based on data analysis provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A risk event analysis method based on data analysis, characterized in that, Includes the following steps: Risk Analysis: Acquire historical risk data sets and current event texts, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event texts, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data consists of text records of various risk events that have occurred, and current event texts are the target event texts to be analyzed. Fusion Analysis: Based on the index of historical risk data, historical related data associated with the current event is obtained from the historical risk data. Based on the risk analysis hot word set, the current event text and historical related data are used to determine the risk hazards. Based on the risk hazard determination results, the risk level is quantitatively scored, and the risk warning is issued based on the risk level quantitative scoring results.

2. The risk event analysis method based on data analysis according to claim 1, characterized in that, Risk analysis includes the following steps: Index building: Import historical risk data into the full-text search engine and build a structured index using the indexing function of the full-text search engine; Data preprocessing: The current event text is cleaned to remove noise data, including stop words and punctuation marks, to obtain the preprocessed text; Text segmentation: Based on a relevant domain dictionary, the preprocessed text is segmented using a segmentation algorithm to break down continuous text content into independent lexical units, resulting in segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the field to which the risk event belongs. Keyword extraction: Core high-frequency keywords that are indicative of risk events were screened from the word segmentation results using the TF-IDF statistical method; Hot keyword collection construction: Integrate and classify core high-frequency keywords to form a structured hot keyword collection for risk analysis.

3. The risk event analysis method based on data analysis according to claim 1, characterized in that, When using the TF-IDF statistical method to screen out core high-frequency keywords that are indicative of risk events from the word segmentation results, the word frequency and inverse document frequency of each word unit after word segmentation are calculated. The word frequency and inverse document frequency of each word unit are combined and calculated. Based on the calculation results, core high-frequency keywords that are highly indicative of risk events are screened out. Among them, word frequency represents the frequency of a word unit in the current text, and inverse document frequency represents the inverse frequency of a word unit in the historical risk dataset.

4. The risk event analysis method based on data analysis according to claim 1, characterized in that, Fusion analysis includes the following operations: Historical correlation data mining: Using the information of personnel related to the current event as the correlation key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical correlation data, and the cumulative frequency of occurrence of historical correlation data is counted; Risk and hazard assessment: Based on the risk hot word set, the current event text and historical related data are compared and analyzed. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hazards in the current event. Risk data weighting percentage calculation: The percentage of total risk data to sample data is calculated to obtain the risk data weighting percentage. Risk data includes current event text identified as risk data and historical related data identified as risk data. Sample data includes current event text and historical related data. Comprehensive quantitative scoring: Based on the proportion of risk data, scores are calculated according to preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the corresponding score is added; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the corresponding score is added; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the corresponding score is added. Classification and early warning output: The risk level is determined according to the cumulative score and a classification and early warning is executed. If the final score exceeds the predetermined value, the early warning information of historical related data is output simultaneously while the current event risk warning is output.

5. A risk event analysis system based on data analysis, characterized in that, It includes a risk analysis module and a fusion analysis module; The risk analysis module is used to perform the following: acquire historical risk data sets and current event texts, establish an index for historical risk data, extract core high-frequency keywords with risk indication significance based on current event texts, and construct a risk analysis hot word set based on core high-frequency keywords. Among them, historical risk data are text records of various risk events that have occurred, and current event texts are target event texts to be analyzed. The fusion analysis module is used to perform the following: based on the index of historical risk data, it retrieves historical related data associated with the current event from the historical risk data; based on the risk analysis hot word set, it determines the risk hazards of the current event text and historical related data; based on the risk hazard determination results, it performs risk level quantification scoring; and based on the risk level quantification scoring results, it provides risk warnings.

6. The risk event analysis system based on data analysis according to claim 5, characterized in that, The risk analysis module is used to perform the following operations: Index building: Import historical risk data into the full-text search engine and build a structured index using the indexing function of the full-text search engine; Data preprocessing: The current event text is cleaned to remove noise data, including stop words and punctuation marks, to obtain the preprocessed text; Text segmentation: Based on a relevant domain dictionary, the preprocessed text is segmented using a segmentation algorithm to break down continuous text content into independent lexical units, resulting in segmentation results. The relevant domain dictionary is a professional vocabulary dictionary related to the field to which the risk event belongs. Keyword extraction: Core high-frequency keywords that are indicative of risk events were screened from the word segmentation results using the TF-IDF statistical method; Hot keyword collection construction: Integrate and classify core high-frequency keywords to form a structured hot keyword collection for risk analysis.

7. The risk event analysis system based on data analysis according to claim 5, characterized in that, When using the TF-IDF statistical method to screen out core high-frequency keywords that are indicative of risk events from the word segmentation results, the word frequency and inverse document frequency of each word unit after word segmentation are calculated. The word frequency and inverse document frequency of each word unit are combined and calculated. Based on the calculation results, core high-frequency keywords that are highly indicative of risk events are screened out. Among them, word frequency represents the frequency of a word unit in the current text, and inverse document frequency represents the inverse frequency of a word unit in the historical risk dataset.

8. The risk event analysis system based on data analysis according to claim 5, characterized in that, The fusion analysis module is used to perform the following operations: Historical correlation data mining: Using the information of personnel related to the current event as the correlation key, based on the index of historical risk data, historical risk data related to the current event is screened and mined from the historical risk data as historical correlation data, and the cumulative frequency of occurrence of historical correlation data is counted; Risk and hazard assessment: Based on the risk hot word set, the current event text and historical related data are compared and analyzed. The comparison and analysis determines whether the current event text and historical related data contain risk hot words. Based on the comparison and analysis results, it is determined whether there are potential risks and hazards in the current event. Risk data weighting percentage calculation: The percentage of total risk data to sample data is calculated to obtain the risk data weighting percentage. Risk data includes current event text identified as risk data and historical related data identified as risk data. Sample data includes current event text and historical related data. Comprehensive quantitative scoring: Based on the proportion of risk data, scores are calculated according to preset scoring rules. The scoring rules are as follows: if the current event text is judged to be risk data, the corresponding score is added; if the frequency of occurrence of historical related data is higher than the preset frequency threshold, the corresponding score is added; if the proportion of risk data in historical related data is higher than the preset proportion threshold, the corresponding score is added. Classification and early warning output: The risk level is determined according to the cumulative score and a classification and early warning is executed. If the final score exceeds the predetermined value, the early warning information of historical related data is output simultaneously while the current event risk warning is output.

Images