Sports event data analysis system and method

Abstract

The invention relates to the technical field of sports events, in particular to a sports event data analysis system and method.The system comprises a real-time data analysis module, a technical action evaluation module, a dynamic competition prediction module and a decision support module, based on a real-time competition data stream, the positions, scores and instant statistical information of players are received, time sequence analysis is executed, and the real-time competition data stream is analyzed; key data changes are recognized, scoring opportunities and defense position points are extracted, and a data pattern analysis result is obtained. According to the invention, through integration of real-time and historical data, accurate capture and deep analysis of key information in a sports event are realized, through instant receiving and processing of player positions, scores and statistical information, a data mode is identified by using time sequence analysis, dynamic understanding of the event is enhanced, more progress data about the event is allowed to be obtained, and the progress of the event is improved. And through comparison of the time convolution network and the action database, the accuracy of technical action evaluation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of sports events, and in particular to a sports event data analysis system and method. Background Art

[0002] A sports event refers to any form of athletic competition, including competition between individuals or teams. The activities can be professional or amateur, ranging from small-scale local competitions to large-scale international events. Sports events can include many types of sports, such as basketball, football, athletics, swimming and tennis, each of which has its own unique rules and competition methods.

[0003] Among them, the sports event data analysis system involves the implementation of a system that can process and analyze sports event-related data. Its main purpose is to provide in-depth analysis by collecting game data (such as player performance, scores, and statistics, etc.) to help coaching teams, players, and analysts optimize training plans and game strategies. In addition, the data analysis system is also used to enhance the audience experience and attract audiences by providing more detailed game analysis and data.

[0004] Existing technologies fail to provide fast enough data processing capabilities to timely analyze and reflect immediate changes in the game, such as key player position movements or scoring. Existing systems lack the ability to effectively combine game video data with real-time statistical information, which limits the overall analysis of the game, reduces the effectiveness of data analysis in providing in-depth game insights, hinders the ability to extract valuable information from the data, affects the decision-making support of coaches and analysts in game preparation and review, and means that available data resources cannot be fully utilized, thereby maximizing the potential value of event data. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a sports event data analysis system and method.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a sports event data analysis system, the system comprising: The real-time data analysis module receives player positions, scores and instant statistical information based on real-time game data streams, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; The technical action evaluation module collects the game video based on the data pattern analysis result, extracts the frame sequence in the video, uses the temporal convolutional network to analyze the technical actions of the players in each frame, and compares them with the stored action database to evaluate the accuracy and effect of the technical actions and obtain action accuracy evaluation information; The game dynamics prediction module predicts the current game dynamics based on the action accuracy evaluation information and historical game data, continuously optimizes the trend prediction by simulating different game scenarios, and identifies potential scoring opportunities and key defensive points to obtain key trend prediction records; The decision support module analyzes the key moments and player actions in the game based on the key trend prediction records, analyzes the potential impact of differential data points, and converts the prediction results into usable information to obtain the results of the game analysis insights.

[0007] The present invention is improved in that the steps of performing time series analysis are specifically: Based on the real-time game data stream, extract player positions, scores and instant statistical information to build the initial data set; A time series analysis is performed on the initial dataset using the formula: ; Identify changes at key time points and obtain time series analysis results, where: Indicates the current time point The data value of The previous time point The data value of is the periodic factor, Indicates time point The change angle, Indicates time point The random error term.

[0008] The present invention is improved in that the step of obtaining the data pattern analysis result is specifically as follows: Based on the time series analysis results, analyzing the impact of data points on the game process to obtain a process impact data set; Based on the process affecting the data set, key data changes are identified, scoring opportunities and defensive position points are extracted, and data pattern analysis results are obtained.

[0009] The present invention is improved in that the analysis steps of the technical actions in each frame are specifically as follows: Based on the data pattern analysis results, extract each frame from the game video, annotate the player position and action in each frame, and construct a frame sequence data set; The frame sequence dataset is analyzed using a temporal convolutional network, using the formula: ; Get the analysis results of each frame of technical action, where: Indicates time point Player action analysis value, is the convolution kernel weight, Representing time point Player action data, is the adjustment coefficient, is the number of history frames.

[0010] The present invention is improved in that the step of obtaining the action accuracy evaluation information is specifically as follows: According to the technical action of each frame, it is compared with the action database, using the formula: ; Evaluate the accuracy and effect of technical movements and obtain preliminary movement accuracy evaluation results, among which, For the The similarity score between the frame action and the standard action in the action database, For the Frame action eigenvalues, It is the first one in the action database corresponding to the standard action. reference feature values, Features The weight of is the number of features for each action; According to the preliminary action accuracy evaluation results, the execution quality of each action is evaluated, and the actions are classified according to their accuracy to obtain action accuracy evaluation information.

[0011] The present invention is improved in that the step of predicting the current game dynamics is specifically as follows: Based on the action accuracy evaluation information, the players' action success rates, common error types and frequencies are collected and integrated with historical game data to obtain action and game analysis data; Based on the action and match analysis data, various potential trends of the current game are analyzed, including performance trends of teams and players, and the dynamics of the current game are predicted to obtain a game trend prediction result.

[0012] The present invention is improved in that the steps of obtaining the key trend prediction record are specifically as follows: Collect data from various game scenarios, conduct initial trend simulation, and obtain preliminary simulation trend results; The preliminary simulation trend results are refined using the formula: ; Continue to optimize the game trend prediction and obtain accurate prediction data, including: Represents the optimized forecast trend value, is the scene relevance factor, is the original simulation result, is the total number of simulation scenarios, is the adjustment coefficient; Core indicators are extracted from the accurate prediction data to identify potential scoring opportunities and key defensive points, and obtain key trend prediction records.

[0013] The present invention is improved in that the steps of obtaining the results of the competition analysis and insights are specifically as follows: Based on the key trend prediction records, extract the data of key moments of the game and the actions of the players, and analyze the influence of the data to obtain a preliminary analysis record of the influence; Based on the preliminary analysis records of the impact, the potential impact of the difference data points was analyzed using the formula: ; The difference data analysis results are obtained, among which, Indicates the impact index of the difference data, It is The deviation value of the data point, It is with The relevance weight of the data points, is the standard deviation, is the total number of data points; Based on the difference data analysis results, the prediction results are converted into usable information to obtain event analysis insights.

[0014] A sports event data analysis method is provided, wherein the sports event data analysis method is performed based on the sports event data analysis system, and comprises the following steps: S1: Based on the real-time game data stream, it receives player positions, scores and instant statistical information, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; S2: Based on the data pattern analysis results, collect the video stream in the game, extract the frame sequence in the video, use the temporal convolutional network to analyze the technical movements of the players in each frame, and compare them with the stored movement database to obtain movement accuracy evaluation information; S3: using the action accuracy evaluation information and combining it with historical competition data, predicting the current game dynamics, continuously optimizing the trend prediction by simulating different game scenarios, identifying potential scoring opportunities and key defensive points, and obtaining key trend prediction records; S4: Based on the key trend prediction records, analyze key moments and player actions in the game, identify difference data points and potential impacts, and obtain key moment analysis data; S5: Based on the key moment analysis data, refine the analysis process of the game data, evaluate the collected data points and statistical results, and convert the prediction results into usable information to obtain the game analysis insights.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through the integration of real-time and historical data, accurate capture and in-depth analysis of key information in sports events are achieved. By instantly receiving and processing player positions, scores and statistical information, and using time series analysis to identify data patterns, the understanding of event dynamics is enhanced, allowing more data on the progress of the game to be obtained. By comparing the temporal convolutional network and the action database, the accuracy of technical action evaluation is improved, and the accuracy of action recognition is increased, thereby comprehensively evaluating the performance of players and making the use of event data more in-depth and extensive. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A module diagram of a sports event data analysis system is proposed for the present invention; Figure 2 A flowchart for performing time series analysis in the present invention; Figure 3 A flowchart for obtaining the data mode analysis results in the present invention; Figure 4 It is a flow chart of the analysis of the technical actions in each frame of the present invention; Figure 5 is a flow chart for obtaining action accuracy evaluation information in the present invention; Figure 6 A flowchart of predicting the current game dynamics in the present invention; Figure 7 A flowchart for obtaining key trend prediction records in the present invention; Figure 8 This is a flow chart for obtaining the results of the event analysis insights in the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.

[0019] Example See also Figure 1 The present invention provides a technical solution: a sports event data analysis system comprising: The real-time data analysis module receives player positions, scores and instant statistical information based on real-time game data streams, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; The technical action evaluation module collects game videos based on data pattern analysis results, extracts frame sequences in the videos, uses a temporal convolutional network to analyze the players' technical actions in each frame, and compares them with the stored action database to evaluate the accuracy and effect of the technical actions and obtain action accuracy evaluation information; The game dynamics prediction module predicts the current game dynamics based on the action accuracy evaluation information and historical game data. It continuously optimizes the trend prediction by simulating different game scenarios, and identifies potential scoring opportunities and key defensive points to obtain key trend prediction records. The decision support module is based on key trend prediction records, analyzes key moments and player actions in the game, analyzes the potential impact of differential data points, and converts prediction results into usable information to obtain event analysis insights.

[0020] The data pattern analysis results include player movement trajectory, scoring probability, and defensive configuration. Movement accuracy evaluation information includes movement synchronization rate, movement response time, and technical movement accuracy. Key trend prediction records include scene control points, key defensive analysis results, and offense-defense conversion speed. The event analysis insights include adjustment suggestions, team performance evaluation results, and key player influence information.

[0021] See also Figure 2 , the steps to perform time series analysis are as follows: Based on the real-time game data stream, extract player positions, scores and instant statistical information to build the initial data set; Player position data is collected through position tracking sensors in the field. The sensors record the player's position information every second during the game. The score information is obtained in real time by the scoreboard system and converted into standardized processable values. Statistical information includes number of shots, pass success rate, etc. After collection, all information is integrated according to a unified timestamp, and the time synchronization protocol is used to perform time correction on data from different sources to ensure consistency in data sequences. At the same time, redundant or erroneous data, such as duplicate records or abnormal values, is cleared. Finally, the cleaned and corrected data is integrated and stored in a structured data table to construct the initial data set.

[0022] Perform time series analysis on the initial dataset using the formula: ; Identify changes at key time points and obtain time series analysis results, where: Indicates the current time point The data value of The previous time point The data value of is a periodic factor, which is used to simulate the periodic changes of data in time series. Indicates time point The change angle, Indicates time point The random error term of At the point in time ,have Indicates the statistical data value of the previous minute, the change angle of the time point , the value of the periodic factor is calculated as , random error , represents the random fluctuation at that time point, and substitute the value into the formula to calculate ,get: ; This result shows that at the current time point , the statistical data value has increased, showing that certain activities in the game (such as scoring or important events) have led to an increase in data, providing a basis for further analysis and decision-making.

[0023] See also Figure 3 , the specific steps for obtaining the data pattern analysis results are: Based on the time series analysis results, analyze the impact of data points on the game process and obtain the process impact data set; The key event points in the time series data are analyzed, including screening according to the amplitude and direction of data changes, extracting the time points related to the key nodes of the game and their corresponding events, and analyzing the statistical characteristics of each time point, including the fluctuation range and frequency distribution of the data, combined with the dynamic changes of the player's position and the score fluctuations, to build the relationship between the data points and the actual progress of the game, and then deduce the degree of influence of each time point on the progress of the game, and generate a process impact data set.

[0024] Based on the process impact data set, identify key data changes, extract scoring opportunities and defensive position points, and obtain data pattern analysis results; Extract key data that lead to changes in important game nodes, and perform detailed calculations on the magnitude and rate of change of the data, as well as the degree of correlation between adjacent time points, to identify the data characteristics of scoring opportunities and the dynamic changes of key defensive positions. Combined with the distribution characteristics of the actual game venue and player positions, the information is comprehensively organized into data pattern analysis results in a time series.

[0025] See also Figure 4 , the analysis steps of the technical actions in each frame are as follows: Based on the data pattern analysis results, each frame is extracted from the game video, the player position and action in each frame are annotated, and a frame sequence dataset is constructed; First, the video data is segmented by frames. The continuous video stream is decomposed into a static image sequence through video decoding technology. The time interval of each frame is determined by the video frame rate. For example, for a video with 30 frames per second, the time interval of each frame is 0.033 seconds. After the segmentation is completed, the frame image is passed to the player position detection module. The module uses the target detection model to locate the players in the image and records their position information through coordinate annotation. At the same time, the action detection module analyzes the player posture in each frame through image features and extracts the action category from it. The action detection module quantifies the posture parameters into numerical forms such as joint angles and limb movement directions. The data is combined and mapped into player action labels. The labeled information of each frame will be stored as a frame sequence data set.

[0026] Use the temporal convolutional network to analyze the frame sequence dataset using the formula: ; Get the analysis results of each frame of technical action, where: Indicates time point Player action analysis value, is the convolution kernel weight, used for the The importance of the previous frame's motion data in the current motion analysis, Representing time point Player action data, is the adjustment coefficient, which is used to adjust the overall impact of the TCN network on the sensitivity of the player's action continuity. Is the number of historical frames, representing the analysis of the current frame action When the TCN network takes into account The motion data of the frame; There is a video clip containing 100 frames. The motion evaluation values ​​of the first five frames have been obtained for each frame through image processing technology. Now the motion evaluation value of the 100th frame is calculated using the formula, that is, , the convolution kernel weights are , adjustment coefficient , the calculation process is: ; The results show that through the calculation of the temporal convolutional network model, the player action analysis value of the 100th frame is 1.655, and the value of 1 is the standard, which means that the player’s technical action in this frame is very active and its dynamic changes are accurately captured, providing a reliable numerical basis for further action quality analysis.

[0027] See also Figure 5 , the specific steps for obtaining the action accuracy evaluation information are as follows: According to the technical action of each frame, it is compared with the action database and the formula is used: ; Evaluate the accuracy and effect of technical movements and obtain preliminary movement accuracy evaluation results, among which, For the The similarity score between the frame action and the standard action in the action database, For the Frame action eigenvalues, It is the first one in the action database corresponding to the standard action. reference feature values, Features The weight of is the number of features for each action; There is an action feature data point, including 5 features, and the standard reference values ​​of each feature are: ; The corresponding player actual action data is: ; The weights are: ; Substitute into the formula to calculate: ; The result shows that the similarity score of the frame action is 2.3, indicating that the player's technical action is highly consistent with the standard action (value 1), indicating that the player's performance at this moment is close to the technical requirements, providing accurate feedback information for targeted improvement of technical actions in future training.

[0028] Based on the preliminary action accuracy assessment results, the execution quality of each action is evaluated, and the actions are classified according to their accuracy to obtain action accuracy evaluation information; First, the accuracy data of action features are classified and counted, and the feature distribution of each group of actions is grouped using a threshold. The process requires standardization of the accuracy score of each action feature, and a clear hierarchical structure is established based on the accuracy range of the classification. Then, the actions are classified according to their accuracy through cluster analysis, which specifically includes first calculating the Euclidean distance between features to determine the degree of similarity of the actions, and then dividing the actions into three categories of high accuracy, medium accuracy, and low accuracy based on the similarity. By further analyzing the mean and variance of the different grouping features in the classification results, the execution quality of each action is evaluated to generate action accuracy evaluation information.

[0029] See also Figure 6 , the steps to predict the current game dynamics are as follows: Based on the action accuracy evaluation information, the players' action success rate, common error types and frequencies are collected and integrated with historical game data, including past game scores and error rates, to obtain action and game analysis data; Collect the players' action success rates and error types, organize each player's action data into a separate record set, and perform standardization, including detailed classification of action types and time series smoothing of action success rates and error rates. At the same time, call the scoring data and error rate data in historical game records, sort them by game date, and extract relevant statistical indicators, such as average score, score variance, median and extreme values ​​of error rate. Then, integrate the action data and historical game statistics, match the corresponding relationship between action records and game data, ensure that each player's action data is consistently associated with the corresponding game records, and generate action and game analysis data.

[0030] Based on action and match analysis data, analyze various potential trends of the current game, including team and player performance trends, and predict the dynamics of the current game to obtain game trend prediction results; Calculate the distribution of each player's action performance in the current game, including the dynamic changes of their historical success rate and error rate, combined with the data trend of pre-match training, to make a probabilistic estimate of the action trend of the current game. For the entire team, analyze its performance trends in different stages of the game, such as the coordination of actions in the opening stage and the speed of action execution in the midfield stage. By calling the statistical model, summarize the performance trends of each stage and generate the overall game trend prediction results for the team and players.

[0031] See also Figure 7 , the specific steps for obtaining key trend forecast records are: Collect data from various game scenarios, conduct initial trend simulation, and obtain preliminary simulation trend results; Detailed data from a variety of game scenarios are collected, including key statistical indicators of the team (such as scoring rate, number of fouls, defensive efficiency) and performance data of individual players (such as shooting percentage, number of assists, number of steals, etc.). After the data is classified by scenario, the parameters are normalized to ensure the comparability of data from different game scenarios. In the normalized data, scoring opportunities, player performance fluctuations, and action changes at critical moments are classified. Subsequently, the classified data is continuously calculated and logically deduced, and the main characteristics and trend changes of the scenario trend data are determined through multi-dimensional correlation operations. Finally, after completing multiple calculations, the preliminary simulation trend results are generated.

[0032] The preliminary simulation trend results are refined using the formula: ; Continue to optimize the game trend prediction and obtain accurate prediction data, including: Represents the optimized predicted trend value, which is used to indicate the trend of the game in various simulation scenarios. For the scenario relevance factor, a weight is set for each simulation scenario to indicate the importance of the scenario in the final prediction. is the original simulation result, which means The predicted trend obtained under each scenario is is the total number of simulation scenarios, is the adjustment coefficient, which is used to adjust the overall sensitivity of the impact of all simulation results; There are 3 scenes, and the original prediction score for each scene is The weight factors are , adjustment coefficient , substitute into the formula to calculate: ; The result shows that the optimized forecast trend is 1.17, which shows a fine-tuned forecasting method that integrates multiple scenario weights and adjustment coefficients, providing reliable data support for trend analysis.

[0033] Extract core indicators from accurate forecast data, identify potential scoring opportunities and key defensive points, and obtain key trend forecast records; First, determine the criteria for scoring opportunities, including field position advantages, players' best status, and opponents' defensive weaknesses. By extracting parameter information of factors, they are weighted according to the classification weights of scoring potential and defensive impact. The key defensive points are identified by regional division, and the field is divided into multiple areas. The key defensive areas are determined according to the density of players' activity trajectories and the opponent's scoring success rate. The information is comprehensively determined and key trend prediction records are generated through sorting and extraction mechanisms.

[0034] See also Figure 8 , the specific steps for obtaining the results of the event analysis insights are as follows: Based on the key trend prediction records, extract the data of key moments of the game and player actions, analyze the influence of the data, and obtain a preliminary analysis record of the influence; Extract and analyze data on key moments and player actions in the game. This process extracts player positions, movement trajectories, and contact points with the ball in sync with time by processing the game video frame by frame. Combined with key event labels on the timeline, such as offensive advancement, scoring attempts, defensive counterattacks, etc., the action types in the extracted data are classified, and the frequency and success rate of each category are counted. The impact of the data on the game results is comprehensively analyzed. The correlation between different action types and key nodes of the game is compared through statistical methods, and the influence value of each action in the game is calculated to generate a preliminary analysis record of the influence; Based on the preliminary analysis records of influence, the potential impact of the difference data points is analyzed using the formula: ; The difference data analysis results are obtained, among which, Indicates the impact index of the difference data, It is The deviation value of the data point, It is with The relevance weight of the data points, is the standard deviation, which is used for normalization to balance the influence of different data points. is the total number of data points; There are three key data points, and the deviation values ​​are , the corresponding relevance weight is , standard deviation Substitute into the formula and calculate: ; The result shows that the comprehensive impact index of the difference data analysis results is 0.5, indicating that after considering various weights and biases, the impact of specific data points on the overall game trend is moderate.

[0035] Based on the difference data analysis results, the prediction results are converted into usable information to obtain the results of event analysis insights; Convert the prediction results into understandable and usable information. First, map the impact index in the difference data with various tactics and key player performance data, analyze the expected scores and defensive effects under different tactical combinations, and organize them into easy-to-interpret tables and charts. Then, annotate the anomalies and key trends in the data to provide clear operational suggestions for coaches and analysis teams. Finally, summarize the converted information to form event analysis insights to guide specific game strategies and real-time adjustments.

[0036] A sports event data analysis method comprises the following steps: S1: Based on the real-time game data stream, it receives player positions, scores and instant statistical information, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; S2: Based on the data pattern analysis results, collect the video stream in the game, extract the frame sequence in the video, use the temporal convolutional network to analyze the technical movements of the players in each frame, and compare them with the stored movement database to obtain movement accuracy evaluation information; S3: Use the action accuracy evaluation information and historical game data to predict the current game dynamics. By simulating different game scenarios, continuously optimize the trend prediction, identify potential scoring opportunities and key defensive points, and obtain key trend prediction records; S4: Based on the key trend prediction records, analyze the key moments and player actions in the game, identify the difference data points and potential impacts, and obtain key moment analysis data; S5: Analyze data based on key moments, refine the analysis process of game data, evaluate the collected data and statistical results, and convert the prediction results into usable information to obtain game analysis insights.

[0037] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A sports event data analysis system, characterized in that: The system comprises: The real-time data analysis module receives player positions, scores and instant statistical information based on real-time game data streams, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; The technical action evaluation module collects the game video based on the data pattern analysis results, extracts the frame sequence in the video, uses the temporal convolutional network to analyze the technical actions of the players in each frame, and compares them with the stored action database to evaluate the accuracy and effect of the technical actions and obtain action accuracy evaluation information; The game dynamics prediction module predicts the current game dynamics based on the action accuracy evaluation information and historical game data, continuously optimizes the trend prediction by simulating different game scenarios, and identifies potential scoring opportunities and key defensive points to obtain key trend prediction records; The decision support module analyzes the key moments and player actions in the game based on the key trend prediction records, analyzes the potential impact of differential data points, and converts the prediction results into usable information to obtain the results of the game analysis insights.

2. The sports event data analysis system according to claim 1, characterized in that: The steps of performing time series analysis are specifically as follows: Based on the real-time game data stream, extract player positions, scores and instant statistical information to build the initial data set; A time series analysis is performed on the initial dataset using the formula: ； Identify changes at key time points and obtain time series analysis results, where: Indicates the current time point The data value of The previous time point The data value of is the periodic factor, Indicates time point The change angle, Indicates time point The random error term.

3. The sports event data analysis system according to claim 1, characterized in that: The steps for obtaining the data mode analysis results are specifically as follows: Based on the time series analysis results, analyzing the impact of data points on the game process to obtain a process impact data set; Based on the process affecting the data set, key data changes are identified, scoring opportunities and defensive position points are extracted, and data pattern analysis results are obtained.

4. The sports event data analysis system according to claim 1, characterized in that: The analysis steps of the technical actions in each frame are specifically as follows: Based on the data pattern analysis results, extract each frame from the game video, annotate the player position and action in each frame, and construct a frame sequence data set; The frame sequence dataset is analyzed using a temporal convolutional network, using the formula: ； Get the analysis results of each frame of technical action, where: Indicates time point Player action analysis value, is the convolution kernel weight, Representing time point Player action data, is the adjustment coefficient, is the number of history frames.

5. The sports event data analysis system according to claim 1, characterized in that: The steps for obtaining the action accuracy evaluation information are specifically as follows: According to the technical action of each frame, it is compared with the action database, using the formula: ； Evaluate the accuracy and effect of technical movements and obtain preliminary movement accuracy evaluation results, among which, For the The similarity score between the frame action and the standard action in the action database, For the Frame action eigenvalues, It is the first one in the action database corresponding to the standard action. reference feature values, Features The weight of is the number of features for each action; According to the preliminary action accuracy evaluation results, the execution quality of each action is evaluated, and the actions are classified according to their accuracy to obtain action accuracy evaluation information.

6. The sports event data analysis system according to claim 1, characterized in that: The steps of predicting the current game dynamics are specifically as follows: Based on the action accuracy evaluation information, the players' action success rates, common error types and frequencies are collected and integrated with historical game data to obtain action and game analysis data; Based on the action and match analysis data, various potential trends of the current game are analyzed, including performance trends of teams and players, and the dynamics of the current game are predicted to obtain a game trend prediction result.

7. The sports event data analysis system according to claim 1, characterized in that: The steps for obtaining the key trend prediction record are specifically as follows: Collect data from various game scenarios, conduct initial trend simulation, and obtain preliminary simulation trend results; The preliminary simulation trend results are refined using the formula: ； Continue to optimize the game trend prediction and obtain accurate prediction data, including: Represents the optimized forecast trend value, is the scene relevance factor, is the original simulation result, is the total number of simulation scenarios, is the adjustment coefficient; Core indicators are extracted from the accurate prediction data to identify potential scoring opportunities and key defensive points, and obtain key trend prediction records.

8. The sports event data analysis system according to claim 1, characterized in that: The specific steps for obtaining the results of the competition analysis and insights are as follows: Based on the key trend prediction records, extract the data of key moments of the game and the actions of the players, and analyze the influence of the data to obtain a preliminary analysis record of the influence; Based on the preliminary analysis records of the impact, the potential impact of the difference data points was analyzed using the formula: ； The difference data analysis results are obtained, among which, Indicates the impact index of the difference data, It is The deviation value of the data point, It is with The relevance weight of the data points, is the standard deviation, is the total number of data points; Based on the difference data analysis results, the prediction results are converted into usable information to obtain event analysis insights.

9. A sports event data analysis method, characterized in that: The sports event data analysis system according to any one of claims 1 to 8 comprises the following steps: Based on real-time game data streams, it receives player positions, scores and instant statistical information, performs time series analysis, identifies key data changes, extracts scoring opportunities and defensive position points, and obtains data pattern analysis results; Based on the data pattern analysis results, the video stream in the game is collected, the frame sequence in the video is extracted, the technical action of the player in each frame is analyzed by using a temporal convolutional network, and the technical action is compared with the stored action database to obtain action accuracy evaluation information; Utilize the action accuracy evaluation information, combined with historical game data, to predict the current game dynamics, continuously optimize trend predictions by simulating different game scenarios, identify potential scoring opportunities and key defensive points, and obtain key trend prediction records; Based on the key trend prediction records, analyze key moments and player actions in the game, identify differential data points and potential impacts, and obtain key moment analysis data; Based on the key moment analysis data, the analysis process of the game data is refined, the collected data points and statistical results are evaluated, and the prediction results are converted into usable information to obtain the game analysis insights.

Images