CN118228889B - Method and system for generating college entrance examination football score based on neural network - Google Patents

Method and system for generating college entrance examination football score based on neural network Download PDF

Info

Publication number
CN118228889B
CN118228889B CN202410579621.3A CN202410579621A CN118228889B CN 118228889 B CN118228889 B CN 118228889B CN 202410579621 A CN202410579621 A CN 202410579621A CN 118228889 B CN118228889 B CN 118228889B
Authority
CN
China
Prior art keywords
data
scoring
performance
model
efficiency
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410579621.3A
Other languages
Chinese (zh)
Other versions
CN118228889A (en
Inventor
杨炳杰
刘添隆
肖建承
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guagnzhou Huaxia Huihai Technology Co ltd
Original Assignee
Guagnzhou Huaxia Huihai Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guagnzhou Huaxia Huihai Technology Co ltd filed Critical Guagnzhou Huaxia Huihai Technology Co ltd
Priority to CN202410579621.3A priority Critical patent/CN118228889B/en
Publication of CN118228889A publication Critical patent/CN118228889A/en
Application granted granted Critical
Publication of CN118228889B publication Critical patent/CN118228889B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • AHUMAN NECESSITIES
    • A63SPORTS; GAMES; AMUSEMENTS
    • A63BAPPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
    • A63B69/00Training appliances or apparatus for special sports
    • A63B69/002Training appliances or apparatus for special sports for football
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/21Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G06F18/2148Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the process organisation or structure, e.g. boosting cascade
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/243Classification techniques relating to the number of classes
    • G06F18/2431Multiple classes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/243Classification techniques relating to the number of classes
    • G06F18/24323Tree-organised classifiers
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • G06N20/20Ensemble learning
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/042Knowledge-based neural networks; Logical representations of neural networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00Computing arrangements using knowledge-based models
    • G06N5/01Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063Operations research, analysis or management
    • G06Q10/0639Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393Score-carding, benchmarking or key performance indicator [KPI] analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • G06Q50/20Education
    • G06Q50/205Education administration or guidance

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Educational Technology (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Medical Informatics (AREA)
  • Primary Health Care (AREA)
  • Electrically Operated Instructional Devices (AREA)

Abstract

The invention relates to the technical field of sports score evaluation, in particular to a method and a system for generating a football score of a middle school examination based on a neural network, which comprises the following steps: based on the athlete's basic data and game performance data, a data envelope analysis model is used to analyze various inputs of the athlete, a random front analysis model is used to process random fluctuations in the data, an efficiency front is constructed, and an efficiency analysis result is generated. According to the invention, through the technologies of fusion data envelope analysis, complex event processing, hypothesis testing, information theory analysis, gradient elevator algorithm, neural symbol learning, ensemble learning and the like, the objectivity, timeliness and accuracy of sports score evaluation are improved. The techniques enable key events in the game to be captured in real time and rapidly update the performance data of the athlete, ensure that the scoring system can instantly reflect the actual performance of the athlete, optimize the scoring criteria by applying hypothesis testing and confidence interval estimation, and enhance the fairness of the scoring process.

Description

Method and system for generating college entrance examination football score based on neural network
Technical Field
The invention relates to the technical field of sports score evaluation, in particular to a method and a system for generating a football score for a middle school examination based on a neural network.
Background
The technical field of sports score evaluation is a field which is focused on quantitative evaluation of sports performance through scientific and technological means. A core goal of this technical field is to provide objective, accurate and efficient assessment methods so that teachers, coaches and assessment staff can fair assess the physical performance of students or athletes. In educational systems, particularly in mid-level educational phases, assessment of athletic performance is an important component of student comprehensive quality assessment involving the mastery of motor skills, physical performance and motor rules. This assessment method needs to consider not only the technical details of sports such as skills and tactics, but also the educational value of sports activities and the impact on student physical and mental development.
The method for generating the football score of the middle school student is an evaluation system specially designed for football items in sports exams of the middle school student, and aims to provide a quantified scoring standard for the performance of the student on the football items. Such a method would cover a number of aspects such as ball skills, performance on a course, tactical understanding capabilities, team cooperation capabilities, etc. In particular, ball winding performance is one of the important evaluation indexes for testing the comprehensive ability of the student to control the ball skill and speed, and the proficiency of the student in controlling the ball skill is quantified by setting standardized winding routes and time limits. The purpose of the generating method is to ensure fairness, objectivity and standardization of the sports performance of the middle school test, so that all examinees are evaluated under the same standard.
While the prior art effectively supports basic assessment of athletic performance, there are significant shortcomings in real-time data processing and analysis. Existing methods face challenges in capturing and resolving critical events that occur in a game, such as ball entry or interception, and in rapidly converting these events into quantifiable scoring criteria. Furthermore, the prior art fails to provide adequate solutions in comprehensively considering athlete performance in many ways, such as comprehensive assessment under the interaction of technical skills and physical fitness. While the prior art has advanced in ensuring standardization and consistency of the assessment process, there are still shortcomings in dynamically adjusting the scoring criteria to reflect the changing performance criteria of athletic sports. In particular, the prior art fails to provide adequate methods and tools in how to adjust the scoring system to ensure that the scoring criteria are both fair and reflect the current level of athletic activity, based on recent game data and performance trends. Furthermore, while the prior art has helped in assessing the past and current performance of an athlete, there remains a deficiency in predicting the athlete's future performance potential and providing targeted development advice. In particular, the lack of the ability to analyze athlete long-term data, identify growth trends and potential, and develop more scientific training programs and game strategies for athletes and coaches, using advanced data analysis and machine learning models, has not provided effective support.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a method and a system for generating a college entrance examination football score based on a neural network.
In order to achieve the above purpose, the present invention adopts the following technical scheme: the method for generating the result of the college entrance examination football based on the neural network comprises the following steps,
S1: based on the athlete basic data and the game performance data, adopting a data envelope analysis model to analyze various inputs of the athlete, adopting a random front analysis model to process random fluctuation in the data, constructing an efficiency front, and generating an efficiency analysis result;
s2: based on the efficiency analysis result, capturing a key event which occurs in real time in a match by adopting a complex event processing algorithm, updating efficiency analysis data in combination with event information, reflecting the instant performance of athletes, and generating an updated efficiency analysis result;
S3: based on the updated efficiency analysis result, adopting a hypothesis testing method, analyzing consistency and fairness in a scoring process, determining reliability of a scoring standard by using confidence interval estimation, optimizing the scoring standard, and generating a corrected scoring standard;
S4: based on the corrected scoring standard, evaluating the information quantity of each competition index by adopting an entropy calculation method in an information theory, analyzing and selecting an index with the greatest influence on a scoring result through information gain, and adjusting a scoring system according to the information quantity, so as to generate an optimized scoring system;
S5: based on the optimized scoring system, a gradient elevator algorithm is adopted, a model is established by utilizing historical data, future performance and growth potential of athletes are predicted, and performance prediction models are generated by cross verification of optimized model parameters;
S6: based on the performance prediction model, a neural symbol learning method is adopted, the action and strategy of football is expressed through symbol logic, the relation between symbols and athlete performances is analyzed by using a deep learning model, and a symbolized skill and strategy evaluation result is generated;
S7: based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, an integrated learning method is adopted to construct an evaluation framework, a score generation and development scheme is provided for athletes, and a comprehensive performance evaluation result and a development plan are generated.
The improvement of the invention is that the efficiency analysis result comprises the comprehensive score, the efficiency rank and the key performance index of the athlete, the updated efficiency analysis result comprises the adjusted athlete efficiency score, the real-time competition performance update and the key event response evaluation, the corrected scoring standard comprises the newly set scoring threshold value, the adjusted skill weight and the scoring dimension, the optimized scoring system comprises the optimized scoring index, the newly added performance dimension and the removed redundancy index, the performance prediction model comprises a predicted growth track, a key capacity improvement area and a potential performance bottleneck, the symbolized skill and strategy evaluation result comprises symbolized skill mastery level, strategy execution effect and personal improvement scheme, and the comprehensive performance evaluation result and development scheme comprise personal comprehensive capacity evaluation, targeted development scheme and formulated training targets.
The improvement of the invention is that, based on the fundamental data and the competition performance data of the athlete, a data envelope analysis model is adopted to analyze various inputs of the athlete, a random fluctuation in the data is processed by a random front analysis model, an efficiency front is constructed, the specific steps for generating an efficiency analysis result are as follows,
S101: based on athlete basic data and game performance data, performing data cleaning by adopting a Pandas library of Python, removing missing values by using a dropna function, filling the missing values by using a fillna function, performing data standardization by using a STANDARDSCALER function, and generating a preprocessed data set;
s102: based on the preprocessed data set, adopting a data envelope analysis model, performing efficiency analysis by using a PyDEA library of Python, setting input parameters and output parameters, calling solve functions to calculate the efficiency value of each athlete, and generating a DEA efficiency analysis result;
S103: based on the DEA efficiency analysis result, a random front edge analysis model is adopted, a front packet of R software is used for processing random fluctuation, model parameters including boundary types and distribution assumptions are set, SFA analysis commands are operated, efficiency values are adjusted, and the randomness of data is referred to, so that an efficiency analysis result is generated.
The invention improves, based on the efficiency analysis result, adopts complex event processing algorithm to capture the key event which happens in real time in the competition, and combines event information to update efficiency analysis data, reflects the instant performance of athletes, generates the updated efficiency analysis result as follows,
S201: based on the efficiency analysis result, monitoring a real-time data stream through an Esper technology of a Java platform, defining event pattern matching logic to identify key events in a match, filtering and matching the match data stream through setting event types and associated attributes, and analyzing the captured event data into structural information immediately to generate a key event information set;
S202: based on the key event information set, carrying out event data analysis and efficiency data updating, analyzing the event information through a Pandas library of Python, including extracting event types, time stamps and player IDs, merging the event data with the original efficiency analysis data through a merge according to the player IDs, and referring to the instantaneity of the event, so that the data of each player can reflect the latest game performance, and generating an event updated data set;
S203: based on the event updated data set, adopting a PyDEA library in Python and a front package of R software again, carrying out data envelope analysis and random front analysis on the updated data, setting input parameters and output parameters of a model again, adjusting parameters and running analysis commands, comprehensively referring to instant performance in a match, reflecting the efficiency state of each athlete, and generating an updated efficiency analysis result.
The invention improves, based on the updated efficiency analysis result, adopts a hypothesis test method, analyzes consistency and fairness in the scoring process, adopts confidence interval estimation to determine the reliability of the scoring standard, optimizes the scoring standard, generates the corrected scoring standard as follows,
S301: based on the updated efficiency analysis result, executing ttest _ind function through SciPy library of Python, setting significance level alpha as 0.05 for two groups of athlete efficiency score data sets through specified parameters, and comparing mean value differences of two independent samples to generate statistical results of hypothesis test;
S302: based on the statistical result of the hypothesis test, using a norm/interval function of SciPy library, wherein parameters comprise a confidence level of 95% and a mean value and a standard deviation calculated based on sample data, providing a range estimation for the efficiency score of the athlete, and generating a confidence interval estimation result of the efficiency score;
s303: and carrying out optimization operation of the scoring standard based on the statistical result of the hypothesis test and the confidence interval estimation result of the efficiency score, and generating a corrected scoring standard by adjusting the scoring parameter including modifying the weight and the threshold value in the scoring standard.
The invention improves, based on the corrected scoring standard, adopts an entropy calculation method in the information theory to evaluate the information quantity of each competition index, selects the index with the largest influence on the scoring result through information gain analysis, adjusts the scoring system according to the information gain analysis, generates the optimized scoring system as follows,
S401: based on the corrected scoring standard, performing entropy calculation on the competition indexes by using a feature_selection module of SciKit-Learn library through Python, evaluating the information quantity of each index, identifying the indexes playing a key role in distinguishing the player performance, and generating an entropy calculation result of the competition indexes;
S402: based on the entropy calculation result of the competition index, an information gain analysis method is adopted, the entropy difference before and after the calculation index is removed, and the information gain analysis result is generated by selecting the index with the largest influence on the scoring result and identifying the most information value in the scoring system;
S403: and adjusting a scoring system based on the information gain analysis result, and updating scoring system configuration, including optimizing index weights and adjusting scoring dimensions, by referring to the contribution of the indexes with the information gain to the accuracy of the scoring system, so that the scoring system fully reveals the real capability level of the athlete, and an optimized scoring system is generated.
The invention is improved in that, based on the optimized scoring system, a gradient elevator algorithm is adopted, a model is established by utilizing historical data, the future performance and growth potential of athletes are predicted, the parameters of the optimized model are verified in a crossing way, the specific steps for generating the performance prediction model are as follows,
S501: based on the optimized scoring system, a gradient hoisting algorithm is selected for model initialization, gradientBoostingRegressor types are called through a scikit-learn library of Python, when a model is initialized, the n_ estimators parameter is set to be 100, the number of trees is defined, the learning_rate parameter is set to be 0.1, the learning rate is set, the max_depth parameter is set to be 4, the maximum depth of the tree is selected, and a prediction model basic structure is generated;
S502: based on the prediction model basic structure, adopting a cross_val_score function in scikit-learn library, setting cv parameters to be 5 for 5-fold cross validation, setting scoring parameters to be R2, adopting R2 as a scoring criterion, evaluating the performance of the model on multiple subsets, capturing the average performance index under the configuration of the model parameters, and generating a cross validation performance evaluation result;
S503: based on the cross-validation performance evaluation result, model parameters are adjusted and optimized, GRIDSEARCHCV is used for grid search of the parameters, multiple value ranges are set for the n_ estimators, learning _rate and the max_depth parameters for analysis, and the optimal model parameter combination is captured to generate a performance prediction model.
The improvement of the invention is that based on the performance prediction model, the action and strategy of football is expressed by symbol logic by adopting a neural symbol learning method, the relation between the symbol and the athlete performance is analyzed by utilizing a deep learning model, the specific steps for generating the symbolized skill and strategy evaluation result are as follows,
S601: based on the performance prediction model, constructing a neural symbol learning network, defining a network structure by using TensorFlow and Keras libraries, including setting an input layer to receive symbolized motion and strategy data, adding a plurality of hidden layers and adopting a ReLU as an activation function, and using a softmax function to perform multi-category prediction by an output layer to generate a deep learning network structure;
S602: based on the deep learning network structure, setting batch_size as 32 and epochs as 100, training a network by using a fit method, optimizing network weight, and generating a trained deep learning model;
s603: based on the trained deep learning model, a predict method is utilized to predict a test set, the output of the deep learning model is converted into scores for the skills and strategies of the athlete, skill and strategy feedback is provided for the athlete, and symbolic skill and strategy evaluation results are generated.
The invention improves, based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skill and the strategy evaluation result, adopts the integrated learning method to construct an evaluation framework, provides a score generation and development scheme for athletes, generates the specific steps of the comprehensive performance evaluation result and the development plan as follows,
S701: based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, adopting an integrated learning method, integrating the prediction results of a plurality of models by utilizing VotingClassifier types in scikit-learn libraries, setting the weight distribution of a plurality of models, providing an evaluation and development plan for athletes, and generating a comprehensive evaluation framework;
S702: based on the comprehensive evaluation frame, refining and optimizing model configuration, and adjusting the weight of a single model in the integrated model according to a prediction result and current feedback to generate an optimized comprehensive evaluation frame;
S703: based on the optimized comprehensive evaluation framework, an integrated learning method is applied, and the comprehensive performance evaluation result and development plan are generated by finely analyzing the performance data of athletes in multiple dimensions and comprehensively reading and predicting the data.
The system comprises a data preprocessing module, an efficiency analysis module, an event processing module, a hypothesis testing module, a scoring system optimizing module, a performance predicting module, a skill strategy evaluating module and a comprehensive evaluation building module;
The data preprocessing module adopts a data cleaning algorithm based on the basic data and the competition performance data of athletes, uses Pandas library of Python to execute dropna and fillna functions to remove missing values and fill up, applies STANDARDSCALER functions to perform data standardization processing, and generates a preprocessed data set;
The efficiency analysis module executes efficiency analysis by using PyDEA libraries of Python based on the preprocessed data set by adopting a data envelope analysis algorithm, sets training time and physical quality as input parameters, uses shooting accuracy and running speed as output parameters, calls solve functions to calculate the efficiency value of each athlete, and generates a DEA efficiency analysis result;
The event processing module monitors real-time data flow by adopting a complex event processing algorithm based on a DEA efficiency analysis result and through an Esper technology of a Java platform, defines event pattern matching logic to identify a key event, analyzes captured event data into structured information and generates a key event information set;
The hypothesis testing module executes ttest _ind function through SciPy library of Python based on key event information set by adopting hypothesis testing algorithm, compares average difference of player efficiency scores, evaluates consistency and fairness of scoring process, and applies norm.interval function to carry out confidence interval estimation to generate corrected scoring standard;
The scoring system optimization module adopts entropy calculation and information gain analysis methods in the information theory based on the corrected scoring standard, evaluates and selects a competition index with the greatest influence on the scoring result by using a feature_selection module of SciKit-Learn library through Python, adjusts the scoring system according to the competition index, and generates an optimized scoring system;
The performance prediction module is used for constructing a prediction model through a scikit-learn library of Python based on an optimized scoring system by adopting a gradient hoisting algorithm, optimizing model parameters by using a cross verification method, predicting the future performance and growth potential of athletes, and generating a performance prediction model;
the skill strategy evaluation module adopts a neural symbol learning method based on a performance prediction model, uses TensorFlow and Keras libraries to construct a deep learning network, analyzes the relation between symbolized motion actions and strategies and player performances, provides evaluation for player skills and strategies, and generates symbolized skill and strategy evaluation results;
The comprehensive evaluation construction module integrates the prediction results of a plurality of models through scikit-learn library of Python by adopting an integrated learning method based on DEA efficiency analysis results, key event information sets, corrected scoring standards, optimized scoring systems, performance prediction models, symbolized skills and strategy evaluation results, provides score generation and development schemes for athletes, and generates comprehensive performance evaluation results and development plans.
Compared with the prior art, the invention has the advantages and positive effects that:
According to the invention, through the technologies of fusion data envelope analysis, complex event processing, hypothesis testing, information theory analysis, gradient elevator algorithm, neural symbol learning, ensemble learning and the like, the objectivity, timeliness and accuracy of sports score evaluation are improved. These techniques enable key events in the game to be captured in real time and the athlete's performance data to be updated quickly, ensuring that the scoring system can reflect the athlete's actual performance instantaneously. In addition, by optimizing the scoring criteria using hypothesis testing and confidence interval estimation, the consistency and fairness of the scoring process is enhanced, while the application of information theory analysis ensures the comprehensiveness and sensitivity of the scoring system. The adoption of the performance prediction model and the neural symbol learning method provides a dedicated development strategy based on deep analysis for athletes, and promotes accurate improvement of skills and deep mining of potential. The integrated learning method is introduced, a plurality of assessment tools and models are further fused, an omnibearing performance assessment and development planning framework is constructed, and more scientific, comprehensive and targeted assessment and development schemes are brought for athletes, coaches and assessment staff.
Drawings
FIG. 1 is a flow chart of a method for generating a performance of a college entrance examination football based on a neural network;
fig. 2 is a schematic diagram of a refinement flow of step S1 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 3 is a schematic diagram of a refinement flow of step S2 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 4 is a schematic diagram of a refinement flow of step S3 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 5 is a schematic diagram of a refinement flow of step S4 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 6 is a schematic diagram of a refinement flow of step S5 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 7 is a schematic diagram of a refinement flow of step S6 in the method for generating a result of a college entrance examination football based on a neural network;
Fig. 8 is a schematic diagram of a refinement flow of step S7 in the method for generating a result of a college entrance examination football based on a neural network;
fig. 9 is a block diagram of a training football score generating system based on a neural network.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," etc. indicate or are based on the orientation or positional relationship shown in the drawings, merely to facilitate description of the invention and simplify description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Examples
Referring to fig. 1, the present invention provides a technical solution: the method for generating the result of the college entrance examination football based on the neural network comprises the following steps,
S1: based on the athlete basic data and the game performance data, adopting a data envelope analysis model to analyze various inputs of the athlete, adopting a random front analysis model to process random fluctuation in the data, constructing an efficiency front, and generating an efficiency analysis result;
S2: based on the efficiency analysis result, capturing a key event which occurs in real time in a match by adopting a complex event processing algorithm, updating efficiency analysis data in combination with event information, reflecting the instant performance of athletes, and generating an updated efficiency analysis result;
s3: based on the updated efficiency analysis result, adopting a hypothesis testing method, analyzing consistency and fairness in the scoring process, determining the reliability of the scoring standard by using confidence interval estimation, optimizing the scoring standard, and generating a corrected scoring standard;
S4: based on the corrected scoring standard, evaluating the information quantity of each competition index by adopting an entropy calculation method in an information theory, analyzing and selecting an index with the greatest influence on a scoring result through information gain, and adjusting a scoring system according to the information quantity, so as to generate an optimized scoring system;
S5: based on the optimized scoring system, a gradient elevator algorithm is adopted, a model is established by utilizing historical data, future performance and growth potential of athletes are predicted, and performance prediction models are generated by cross verification of optimized model parameters;
s6: based on a performance prediction model, a neural symbol learning method is adopted, the action and strategy of football is represented through symbol logic, the relation between symbols and athlete performances is analyzed by using a deep learning model, and a symbolized skill and strategy evaluation result is generated;
S7: based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, an integrated learning method is adopted to construct an evaluation framework, a score generation and development scheme is provided for athletes, and a comprehensive performance evaluation result and a development plan are generated.
The efficiency analysis results comprise comprehensive scores, efficiency ranks and key performance indexes of athletes, the updated efficiency analysis results comprise adjusted athlete efficiency scores, real-time game performance updates and key event response evaluations, the corrected scoring standards comprise newly set scoring thresholds, adjusted skill weights and scoring dimensions, the optimized scoring system comprises optimized scoring indexes, newly added performance dimensions and removed redundant indexes, the performance prediction model comprises predicted growth tracks, key capacity improvement areas and potential performance bottlenecks, the symbolized skill and strategy evaluation results comprise symbolized skill mastery levels, strategy execution effects and personal improvement schemes, and the comprehensive performance evaluation results and development plans comprise personal comprehensive capacity evaluation, targeted development schemes and formulated training targets.
In step S1, the performance of the athlete is evaluated by using a data envelope analysis model (DEA) and a random front analysis model (SFA) in combination. First, the athlete's basic data and game performance data are collected, and the data format includes, but is not limited to, excel or CSV files, in which key indexes such as training time, physical quality test result, performance in the game, etc. of the athlete are recorded in detail. The data is then pre-processed, including cleaning of missing values and normalization, using the Python programming language, with the Pandas library to ensure data quality and consistency. Subsequently, DEA analysis is performed using the PyDEA library, which calculates the position of each athlete with respect to the efficiency front by setting input (e.g., training time and physical quality) and output parameters (e.g., goal accuracy and running speed), thereby obtaining a preliminary efficiency score. To further improve the accuracy of the assessment, the SFA analysis is performed using the front package in the R package with the DEA result as input, which adjusts the efficiency score by processing random fluctuations in the data to more accurately reflect the athlete's actual performance. Finally, the series of operations generate report files containing analysis results of the comprehensive efficiency of the athletes, and scientific basis is provided for subsequent scoring and development planning.
In step S2, key events in the game are captured in real time using a Complex Event Processing (CEP) algorithm, and efficiency analysis data of the athlete is updated in conjunction with the event information. This step begins by collecting real-time data during the game, including player position, ball movement trajectory, etc., through real-time data stream capture techniques such as sensors and video analysis tools. Then, event patterns and logic are defined by using the CEP software platform, such as Esper, to identify key events in the game in real time, such as goal, boost, key defender, etc. Whenever these events are identified, the system automatically combines the event information with the existing efficiency analysis data, using the Pandas library of Python for data consolidation and updating, ensuring that each athlete's data can reflect their immediate performance in the game in real time. In this way, updated efficiency analysis results are generated that include not only the athlete's base performance data, but also integrate real-time performance in the game, providing data support for more accurate assessment of the athlete's overall performance.
In step S3, based on the updated efficiency analysis result, the consistency and fairness of the scoring process are analyzed by using a hypothesis testing method, and meanwhile, the reliability of the scoring standard is determined by using confidence interval estimation. Firstly, carrying out statistical analysis by adopting SciPy libraries of Python, executing t-test and other hypothesis test functions, and comparing the mean value differences of the efficiency scores of different groups of athletes, thereby evaluating the consistency of the scoring process. Next, using the confidence interval estimation method, a 95% confidence interval for the score is calculated based on the sample mean and standard deviation of the player efficiency score, which helps evaluate the stability and reliability of the scoring criteria. Finally, according to the statistical analysis result, the scoring process is optimized by adjusting the weight and the threshold value in the scoring standard, so that the fairness and the accuracy of scoring are ensured. Through the refinement operations, a corrected scoring standard is generated, and a solid statistical basis is provided for fairly and objectively evaluating the performance of athletes.
In step S4, the scoring system is optimized by the entropy calculation method of the information theory and the information gain analysis. Firstly, based on the corrected scoring standard, performing entropy calculation on the competition index by using a feature_selection module of SciKit-Learn library of Python. The process includes quantifying the amount of information for each game indicator, such as shot accuracy, running speed, and an indicator with a high entropy value indicates that the indicator has a large uncertainty and amount of information in distinguishing player performance. Then, through information gain analysis, the variation of the information entropy before and after considering a certain index is calculated, and the index with the greatest influence on the scoring result is identified. These indicators reflect key aspects of athlete performance, and are critical to the accuracy and fairness of the scoring system. Finally, based on the results of entropy calculation and information gain analysis, adjusting a scoring system, including optimizing index weights and adjusting scoring dimensions, to generate an optimized scoring system. The process ensures that the scoring system can more accurately reflect the real performance of the athlete, and improves the objectivity and scientificity of scoring.
In step S5, a gradient hoist algorithm (Gradient Boosting Machine, GBM) is used to predict future performance and growth potential of the athlete. First, historical data is collected and consolidated as a training set of models, including training records of athletes, game achievements, and the like. The model is initialized by calling GradientBoostingRegressor classes through scikit-learn library of Python, key parameters are set such as n_ estimators to define the number of trees, learning_rate is set to learn rate, and max_depth is selected to be the maximum depth of the tree. Next, cross-validation is performed using a cross val score function to evaluate the model's performance on different subsets of data and optimize the model parameters. Finally, a performance prediction model is generated based on the optimal parameter combination. The model can predict the performance trend and potential of athletes in a period of time in the future, and provides scientific basis for developing training and development plans.
In step S6, the action and strategy of football is represented by symbol logic by using a neural symbol learning method, and the relation between the symbol and the player performance is analyzed by using a deep learning model. First, the player's actions and strategies in the game are converted into computer-understandable symbolic logic by symbolizing the representation. Then, a deep learning model is built using TensorFlow and Keras libraries, the model structure including input layers receiving symbolized data, multiple hidden layers processing the data using a ReLU activation function, and output layers employing a softmax function for multi-class prediction. Training the model by a fit method, and predicting the test set by a predict method to generate a symbolized skill and strategy evaluation result. The results provide a deep skill and strategic analysis for coaches and athletes, guiding improvement in athlete skills and tactics.
In the step S7, the efficiency analysis result, the performance prediction model and the symbolized skill evaluation result are comprehensively utilized, and a comprehensive evaluation framework is constructed through an integrated learning method. And (3) integrating the results of a plurality of prediction models by using VotingClassifier types in scikit-learn libraries to comprehensively evaluate the performance of the athlete. The accuracy of the assessment framework is optimized by adjusting the weight of each model in the integration. Finally, the generated comprehensive performance evaluation result and development plan are not only based on the historical performance of the athlete, but also consider the potential growth space and strategy adjustment potential, and provide comprehensive growth and development guidance for the athlete.
Referring to fig. 2, based on athlete's base data and game performance data, various athlete inputs are analyzed using a data envelope analysis model, random fluctuations in the data are processed using a random front analysis model, efficiency fronts are constructed, specific steps for generating efficiency analysis results are as follows,
S101: based on athlete basic data and game performance data, performing data cleaning by adopting a Pandas library of Python, removing missing values by using a dropna function, filling the missing values by using a fillna function, performing data standardization by using a STANDARDSCALER function, and generating a preprocessed data set;
s102: based on the preprocessed data set, adopting a data envelope analysis model, performing efficiency analysis by using a PyDEA library of Python, setting input parameters and output parameters, calling solve functions to calculate the efficiency value of each athlete, and generating a DEA efficiency analysis result;
S103: based on the DEA efficiency analysis result, a random front edge analysis model is adopted, the front package of R software is used for processing random fluctuation, model parameters including boundary types and distribution assumptions are set, SFA analysis commands are operated, efficiency values are adjusted, and the randomness of data is referred to, so that the efficiency analysis result is generated.
In the S101 substep, the basic data and the game performance data of the athlete are preprocessed through the Pandas library of Python, so that the quality of the analyzed basic data is ensured. Specifically, the dropna function is first used to remove any missing values in the dataset, ensuring data integrity. Next, the remaining blank values are filled with fillna functions, filling is performed according to the average value of the previous and subsequent data or specific logic, which ensures continuity and consistency of the dataset. Finally, the data is subjected to standardized processing by using STANDARDSCALER functions, all characteristic values are converted into a format with the mean value of 0 and the standard deviation of 1, so that the influence caused by different amounts of data is eliminated, and the subsequent efficiency analysis is more accurate and fair. The series of operations generates a preprocessed data set that provides clean, normalized input data for data envelope analysis and random front analysis.
In the sub-step S102, the pre-processed dataset is efficiently analyzed by a data envelope analysis model (DEA) using the PyDEA library of Python. Firstly, setting input parameters including basic data such as training time and physical quality, and setting output parameters as competition performance indexes such as shooting accuracy and running speed, and clearly analyzing input and output frames. Subsequently, a solve function is called to perform DEA analysis, and the efficiency value of each athlete is calculated. The DEA model evaluates the relative efficiency of each athlete by comparing the output performance of the athlete under the given input condition, and the generated DEA efficiency analysis result reveals the relative advantages and disadvantages of the athlete performance and provides basis for subsequent analysis and improvement.
In the S103 substep, random fluctuations in the DEA efficiency analysis result are processed by a random front edge analysis model (SFA) using the front package of R software. First, model parameters are set, including boundary types and distribution assumptions, etc., to establish the basic framework of the analytical model. Then, SFA analysis commands are executed to adjust the efficiency values of each player, taking into account the randomness of the data and the influence of external environmental factors. SFA analysis provides more accurate adjustment for the efficiency value of each athlete by finely processing DEA results, and the generated efficiency analysis results are closer to the actual performance of the athlete, thereby providing more scientific basis for evaluation and improvement.
Referring to fig. 3, based on the efficiency analysis result, a complex event processing algorithm is used to capture key events occurring in real time in a game, and update efficiency analysis data in combination with event information, reflect the instant performance of athletes, generate updated efficiency analysis results as follows,
S201: based on the efficiency analysis result, monitoring the real-time data stream through the Esper technology of the Java platform, defining event pattern matching logic to identify key events in the game, filtering and matching the game data stream through setting event types and associated attributes, and analyzing the captured event data into structured information immediately to generate a key event information set;
S202: based on a key event information set, carrying out event data analysis and updating of efficiency data, analyzing the event information through a Pandas library of Python, including extracting event types, time stamps and player IDs, merging the event data with original efficiency analysis data through a merge according to the player IDs, and referring to the instantaneity of the event, so that the data of each player can reflect the latest game performance, and generating an event updated data set;
s203: based on the event updated data set, adopting PyDEA library in Python and front package of R software again, carrying out data envelope analysis and random front analysis on the updated data, setting input parameters and output parameters of the model again, adjusting parameters and running analysis commands, comprehensively referring to instant performance in the competition, reflecting the efficiency state of each athlete, and generating an updated efficiency analysis result.
In the sub-step S201, based on the existing efficiency analysis result, real-time data stream monitoring is implemented by the Esper technology of the Java platform. Esper technology is capable of handling a large number of event data streams and identifying key events in a game, such as goals, yellow cards, etc., by defining complex event processing logic. In this process, event types and associated attributes are set as filtering conditions to ensure that only relevant event data is captured and parsed. The captured data is then parsed according to a predefined structure, converted into a key event information set. The process not only improves the efficiency of data processing, but also provides real-time and accurate data support for subsequent analysis.
In the S202 substep, the captured key event information is resolved in detail by the Pandas library of Python and integrated with the efficiency analysis data. In the process, information such as event type, time stamp, player ID and the like is extracted, and is combined with original efficiency analysis data through a merge function to form an updated data set. The process ensures timeliness and integrity of the data, so that the performance data of each athlete can timely reflect the actual performance of the athlete in the competition, and a basis is provided for subsequent efficiency analysis.
In a sub-step S203, the updated dataset is re-analyzed using the PyDEA library of Python and the front package of R software. By adjusting input and output parameters, such as training time and physical quality are used as input, shooting accuracy and running speed are used as output, and analysis commands are operated, comprehensive efficiency analysis is performed for instant performance in a game. The process aims to more accurately reflect the efficiency state of the athlete by considering the real-time change of the match, and an updated efficiency analysis result is generated. The result not only provides instant athlete performance feedback for coaches, but also provides scientific basis for athlete training and game strategy adjustment.
Referring to fig. 4, based on the updated efficiency analysis result, the consistency and fairness of the scoring process are analyzed by adopting a hypothesis testing method, the reliability of the scoring criteria is determined by using confidence interval estimation, the scoring criteria is optimized, the specific steps for generating the corrected scoring criteria are as follows,
S301: based on the updated efficiency analysis result, executing ttest _ind function through a SciPy library of Python, setting the significance level alpha to 0.05 for the efficiency score dataset of two groups of athletes through the designated parameters, and comparing the mean difference of two independent samples to generate a statistical result of hypothesis test;
S302: based on the statistical result of hypothesis test, using the norm/interval function of SciPy library, the parameters including confidence level 95% and mean and standard deviation calculated based on sample data, to provide a range estimation for athlete efficiency score, to generate confidence interval estimation result of efficiency score;
s303: and carrying out optimization operation of the scoring standard based on the statistical result of the hypothesis test and the confidence interval estimation result of the efficiency score, and generating a corrected scoring standard by adjusting the scoring parameter including modifying the weight and the threshold value in the scoring standard.
In the S301 substep, statistical analysis is performed on the updated efficiency analysis results by performing a t-test independent sample test (ttest _ind function) using SciPy libraries of Python. In this process, the efficiency score dataset of two groups of athletes was designated as input, and the significance level α was set to 0.05, to compare whether there was a significant difference in the mean between the two groups of data. The application of the method aims at checking consistency and fairness of the scoring process and ensuring scientificity and rationality of scoring standards. The generated hypothesis testing statistical result provides a quantitative evaluation basis for the scoring process, and whether deviation exists in a scoring system is revealed.
In the S302 substep, using the norm/interval function of the SciPy library, a confidence interval for the player efficiency score is calculated based on the statistical result of the hypothesis test. By assigning a confidence level of 95% and inputting the mean and standard deviation calculated based on the sample data, a range estimate of the confidence interval is provided for each athlete's efficiency score. The process is helpful for evaluating the reliability of the scoring standard, provides scientific basis for correcting the scoring standard, and ensures the accuracy and stability of the scoring result.
In the S303 substep, the scoring criteria are optimized by Python script based on the statistical result of the hypothesis test and the confidence interval estimation result of the efficiency score. Including adjusting weights and thresholds in the scoring criteria based on the analysis results to ensure consistency and fairness of the scoring process. Through this optimization operation, a corrected scoring criterion is generated. The process not only improves the scientificity and transparency of the scoring system, but also provides more reasonable and fair evaluation for the athlete, and ensures that the scoring result can truly reflect the performance and capability of the athlete.
Referring to fig. 5, based on the corrected scoring criteria, the information amount of each game index is evaluated by adopting an entropy calculation method in the information theory, the index having the greatest influence on the scoring result is selected by information gain analysis, the scoring system is adjusted accordingly, the specific steps for generating the optimized scoring system are as follows,
S401: based on the corrected scoring standard, performing entropy calculation on the competition indexes by using a feature_selection module of SciKit-Learn library through Python, evaluating the information quantity of each index, identifying the indexes playing a key role in distinguishing the player performance, and generating an entropy calculation result of the competition indexes;
S402: based on the entropy calculation result of the competition index, an information gain analysis method is adopted, the difference of the entropy before and after the calculation index is removed, and the information gain analysis result is generated by selecting the index with the greatest influence on the scoring result and identifying the most information value in the scoring system;
S403: based on the information gain analysis result, adjusting the scoring system, and updating the scoring system configuration, including optimizing the index weight and adjusting the scoring dimension, referring to the contribution of the index with the information gain to the accuracy of the scoring system, so that the scoring system fully reveals the real capability level of the athlete, and generating an optimized scoring system.
In the S401 substep, information entropy calculation is performed on the game index under the corrected scoring standard by using the feature_selection module in the SciKit-Learn library of Python. The process involves taking as input data each of the game metrics (e.g., training time, physical attributes, goal accuracy, and running speed), formatted as a structured data table, where each row represents a record of one athlete, and each column corresponds to one metric. By calculating the entropy of the information of these indices, we can evaluate the amount of information of each index in distinguishing the athlete's performance, i.e. its contribution to predicting the athlete's performance. The index of high information entropy indicates that the index has higher value in distinguishing different athlete performances. The generated entropy calculation result helps to identify key performance indexes, and provides basis for subsequent scoring system optimization.
The S402 substep adopts an information gain analysis method, and calculates the information gain of each index through a Python script. The process involves comparing the information entropy differences of the system with and without a certain index to evaluate the contribution of introducing the index to the improvement of the scoring system. The index with the greatest information gain is selected as the key index in the scoring system, as it contributes most to distinguishing athlete performance. The information gain analysis result generated by the step guides the further optimization of the scoring system, and ensures that the scoring system can reflect the actual performance of the athlete more accurately.
S403, adjusting a scoring system through the Python script, wherein the step comprises the steps of optimizing the weight of the scoring index and adjusting the scoring dimension. In operation, the weights of the indicators in the scoring system are adjusted and the scoring dimensions are reconfigured according to the indicators with high information gain. The purpose of this process is to ensure that the scoring system is able to evaluate the performance of the athlete more comprehensively and equitably, and is particularly effective in distinguishing between high and low performing athletes. The generated optimized scoring system not only improves the accuracy and fairness of scoring, but also provides more valuable feedback and development advice for coaches and athletes by more carefully reflecting each ability of the athlete.
Referring to fig. 6, based on the optimized scoring system, a gradient hoisting algorithm is adopted to build a model by using historical data, predict the future performance and growth potential of the athlete, and by cross-verifying the optimized model parameters, the specific steps for generating the performance prediction model are as follows,
S501: based on the optimized scoring system, selecting a gradient hoisting algorithm to initialize a model, calling GradientBoostingRegressor types through a scikit-learn library of Python, setting an n_ estimators parameter as 100 to define the number of trees, setting a learning_rate parameter as 0.1 to set a learning rate, and setting a max_depth parameter as 4 to select the maximum depth of the trees to generate a prediction model base structure;
S502: based on a prediction model infrastructure, adopting a cross_val_score function in scikit-learn library, setting cv parameters to 5 for 5-fold cross validation, setting scoring parameters to R2, adopting R2 as a scoring criterion, evaluating the performance of the model on multiple subsets, capturing average performance indexes under the parameter configuration of the model, and generating a cross validation performance evaluation result;
s503: based on the cross-validation performance evaluation result, the model parameters are adjusted and optimized, the GRIDSEARCHCV is utilized to conduct grid search of the parameters, various value ranges are set for the n_ estimators, learning _rate and the max_depth parameters for analysis, and the optimal model parameter combination is captured to generate the performance prediction model.
In a sub-step S501, the model is initialized by a gradient hoist algorithm, which uses class GradientBoostingRegressor in the scikit-learn library of Python. First, the data format is a structured data set containing historical performance and related characteristics of the athlete, each row representing a record of the athlete, and the columns containing metrics such as training time, physical quality, goal accuracy, and running speed. At model initialization, setting n_ estimators =100 defines the number of decision trees in the model, learning_rate=0.1 determines the contribution in each tree training process, and max_depth=4 limits the maximum depth of the tree, preventing model overfitting. The process builds a basic predictive model framework aimed at predicting a player's future performance based on its historical data.
In a substep S502, model performance is assessed by a 5-fold cross-validation method, using the cross val score function of the scikit-learn library. The function divides the dataset into 5 different subsets, trains using 4 subsets at a time and tests the model performance on the remaining one subset, repeating this process 5 times. The scoring = R2 parameter specifies the use of R2 as a scoring criterion to quantify the accuracy of the model predictive capability. The cross-validation performance evaluation result generated by the step reveals the average performance of the model on different data subsets, and provides basis for further adjustment of model parameters.
In a substep S503, optimization of model parameters is performed based on the results of the cross-validation, and a parametric grid search is performed using GRIDSEARCHCV. The process involves systematically testing different combinations of parameters within a predetermined parameter range (e.g., different values of n_ estimators, learning _rate, max_depth) to find parameter settings that optimize the model performance. This step determines the model configuration that best improves the prediction accuracy by analyzing the behavior of the model (using R2 scoring criteria) under different parameter combinations. The generated performance prediction model is based on the optimized parameters, has higher prediction accuracy and generalization capability, and can provide more accurate prediction about future performance and growth potential of athletes for coaches.
Referring to fig. 7, based on a performance prediction model, the actions and strategies of football are represented by symbol logic by adopting a neural symbol learning method, the relation between symbols and athlete performances is analyzed by using a deep learning model, the specific steps for generating the symbolic skill and strategy evaluation result are as follows,
S601: based on a performance prediction model, constructing a neural symbol learning network, defining a network structure by using TensorFlow and Keras libraries, including setting an input layer to receive symbolized motion and strategy data, adding a plurality of hidden layers and adopting a ReLU as an activation function, and using a softmax function to perform multi-category prediction by an output layer to generate a deep learning network structure;
S602: based on a deep learning network structure, setting batch_size as 32 and epochs as 100, training a network by using a fit method, optimizing network weight, and generating a trained deep learning model;
s603: based on the trained deep learning model, a predict method is utilized to predict a test set, the output of the deep learning model is converted into scores for the skills and strategies of the athlete, skill and strategy feedback is provided for the athlete, and symbolic skill and strategy evaluation results are generated.
In the sub-step S601, a neural symbol learning network is constructed, and TensorFlow and Keras libraries are used to define a network structure, so as to process symbolic data of actions and strategies of football sports. The data format is a multi-dimensional array, each element representing a symbolized action or strategic feature of the player, such as ball transfer direction, player movement speed, etc., encoded as values for processing by the neural network. The network structure includes an input layer for receiving the symbolized data; a plurality of hidden layers, which adopt a ReLU activation function to increase the nonlinear processing capacity of the model; the output layer uses a softmax function to implement multi-class predictions. The design aims to deeply mine complex relations between symbolized data and athlete performances, and a deep learning network structure capable of accurately identifying and evaluating athlete skills and strategies is generated.
In the S602 substep, the neural network is trained using the fit method by setting the batch size (batch_size) to 32 and the number of iterations (epochs) to 100. In this process, the model minimizes the prediction error by processing the data batch by batch and updating the weights in each iteration. The choice of batch size and number of iterations balances training efficiency and model performance, ensuring that the model can learn patterns in the data sufficiently without overfitting. After training is completed, the generated deep learning model optimizes the network weight, and accuracy of model prediction skills and strategy evaluation is improved.
In a substep S603, a test set is predicted using predict methods based on the optimized deep learning model, and the output of the model is converted into specific scores for athlete skills and strategies. In the process, the model utilizes the relation between the learned symbolized data and the performance to evaluate the unseen data of the athletes, and generates evaluation results of skills and strategies of each athlete. These assessment results provide direct feedback to the athlete regarding their skill and performance of the strategy, helping them to understand which aspects they are performing well in and which aspects need improvement, thereby guiding the training and adjustment of the game strategy.
Referring to fig. 8, based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, the integrated learning method is adopted to construct an evaluation framework, provide the score generation and development scheme for the athlete, generate the comprehensive performance evaluation result and development plan as follows,
S701: based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, adopting an integrated learning method, integrating the prediction results of a plurality of models by utilizing VotingClassifier types in scikit-learn library, setting the weight distribution of a plurality of models, providing an evaluation and development plan for athletes, and generating a comprehensive evaluation framework;
S702: based on the comprehensive evaluation frame, refining and optimizing model configuration, and adjusting the weight of a single model in the integrated model according to the prediction result and current feedback to generate an optimized comprehensive evaluation frame;
s703: based on the optimized comprehensive evaluation framework, an integrated learning method is applied, and the comprehensive performance evaluation result and development plan are generated by carrying out detailed analysis on the performance data of athletes in multiple dimensions and carrying out comprehensive interpretation and prediction on the data.
In the step S701, the prediction results of the different models are integrated by the ensemble learning method to construct a comprehensive evaluation framework. In the process, votingClassifier types in scikit-learn libraries are utilized, the data format is mainly a predictive probability or decision label output by each model, and the data reflect the performance of athletes in various aspects, such as efficiency analysis results, skill and strategy evaluation and the like. By setting weight distribution for different models, the method integrates multidimensional information from efficiency analysis, performance prediction and symbolization skill and strategy evaluation, and improves the accuracy and reliability of evaluation. The execution of this step creates a framework that can provide comprehensive athlete assessment and development advice, providing scientific basis for athlete training and competition strategies.
In the sub-step S702, the weights of the single models in the integrated model are adjusted according to the model prediction result and the real-time feedback by refining and optimizing the model configuration. In the process, the contribution degree of each model to the final evaluation result is considered, the optimization operation is mainly realized through a Python script, and the model weights including a data envelope analysis model, a random front edge analysis model, a performance prediction model and the like are adjusted. The optimized comprehensive evaluation framework reflects the instant performance and potential development trend of the athlete more sensitively, accuracy and practicability of the evaluation result are improved, and the development plan is more targeted and effective.
In the step S703, based on the optimized comprehensive evaluation framework, the integrated learning method is applied to perform deep analysis on the performance data of the athlete in multiple dimensions, and comprehensive interpretation and prediction are performed on the data. In the operation process, various algorithm models are integrated to carry out all-round analysis on the past and present performances of the athlete, and simultaneously, the future development potential of the athlete is predicted. In this way, a development plan is generated that includes athlete comprehensive ability assessment, targeted development advice, and training goals. The achievements not only provide clear lifting direction for the athlete, but also provide scientific training guidance and basis for game strategy establishment for the coach team.
Referring to fig. 9, the system for generating the center-test football score based on the neural network comprises a data preprocessing module, an efficiency analysis module, an event processing module, a hypothesis testing module, a scoring system optimizing module, a performance predicting module, a skill strategy evaluating module and a comprehensive evaluation constructing module;
The data preprocessing module adopts a data cleaning algorithm based on the athlete basic data and the game performance data, uses Pandas library of Python to execute dropna and fillna functions to remove missing values and fill up, and applies STANDARDSCALER functions to perform data standardization processing to generate a preprocessed data set;
The efficiency analysis module executes efficiency analysis by using PyDEA library of Python based on the preprocessed data set, sets training time and physical quality as input parameters, uses shooting accuracy and running speed as output parameters, calls solve function to calculate efficiency value of each athlete, and generates DEA efficiency analysis result;
the event processing module monitors real-time data flow by adopting a complex event processing algorithm based on a DEA efficiency analysis result and through an Esper technology of a Java platform, defines event pattern matching logic to identify a key event, analyzes captured event data into structured information and generates a key event information set;
The hypothesis testing module executes ttest _ind function through SciPy library of Python based on key event information set by adopting hypothesis testing algorithm, compares average difference of player efficiency scores, evaluates consistency and fairness of scoring process, and applies norm.interval function to carry out confidence interval estimation to generate corrected scoring standard;
The scoring system optimization module adopts entropy calculation and information gain analysis methods in the information theory based on the corrected scoring standard, evaluates and selects a competition index with the greatest influence on the scoring result by using a feature_selection module of SciKit-Learn library through Python, adjusts the scoring system according to the competition index, and generates an optimized scoring system;
the performance prediction module is used for constructing a prediction model through a scikit-learn library of Python based on an optimized scoring system by adopting a gradient hoisting algorithm, optimizing model parameters by using a cross verification method, predicting the future performance and growth potential of athletes, and generating a performance prediction model;
The skill strategy evaluation module adopts a neural symbol learning method based on a performance prediction model, uses TensorFlow and Keras libraries to construct a deep learning network, analyzes the relation between symbolized motion actions and strategies and player performances, provides evaluation for player skills and strategies, and generates symbolized skill and strategy evaluation results;
The comprehensive evaluation construction module integrates the prediction results of a plurality of models through scikit-learn library of Python by adopting an integrated learning method based on the DEA efficiency analysis result, the key event information set, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skill and the strategy evaluation result, provides a score generation and development scheme for athletes, and generates a comprehensive performance evaluation result and a development plan.
Through data preprocessing, accuracy and reliability of analysis data are ensured, and high-quality basic data are provided for subsequent efficiency analysis and event processing. And secondly, the efficiency analysis module and the event processing module are applied, so that the performance evaluation of the athlete is more objective and comprehensive, not only basic skills and physical performance are considered, but also real-time performance in the competition is brought into an evaluation range, and the timeliness and the accuracy of the evaluation are enhanced. In addition, the consistency and fairness of the scoring process are further improved by introducing a hypothesis testing and scoring system optimizing module, and the scoring standard is corrected and optimized through a scientific method. The performance prediction module and the skill strategy evaluation module are combined, so that not only is the current performance of the athlete assessed, but also the future development potential of the athlete is predicted, and the skill and the strategy are deeply analyzed, so that guidance is provided for the development of the athlete. And finally, the comprehensive evaluation construction module integrates the results of all modules through an integrated learning method, generates comprehensive performance evaluation results and a targeted development plan, provides powerful decision support for coaches and athletes, greatly improves efficiency and scientificity of generating the football results of the college entrance examination, and is beneficial to promoting comprehensive development of the skills of the athletes and integral improvement of the football training level.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.

Claims (8)

1. The method for generating the college entrance examination football score based on the neural network is characterized by comprising the following steps of:
Based on the athlete's basic data and game performance data, the data envelope analysis model is used to analyze various inputs of the athlete, the random fluctuation in the data is processed by the random front analysis model, the efficiency front is constructed, the specific steps for generating the efficiency analysis result are as follows,
Based on athlete basic data and game performance data, performing data cleaning by adopting a Pandas library of Python, removing missing values by using a dropna function, filling the missing values by using a fillna function, performing data standardization by using a STANDARDSCALER function, and generating a preprocessed data set;
Based on the preprocessed data set, adopting a data envelope analysis model, performing efficiency analysis by using a PyDEA library of Python, setting input parameters and output parameters, calling solve functions to calculate the efficiency value of each athlete, and generating a DEA efficiency analysis result;
based on the DEA efficiency analysis result, adopting a random front edge analysis model, processing random fluctuation by using a front packet of R software, setting model parameters including boundary types and distribution assumptions, running an SFA analysis command, adjusting an efficiency value and referring to randomness of data to generate an efficiency analysis result;
Based on the efficiency analysis result, capturing the key events occurring in real time in the competition by adopting a complex event processing algorithm, updating efficiency analysis data in combination with event information, reflecting the instant performance of athletes, generating the updated efficiency analysis result as follows,
Based on the efficiency analysis result, monitoring a real-time data stream through an Esper technology of a Java platform, defining event pattern matching logic to identify key events in a match, filtering and matching the match data stream through setting event types and associated attributes, and analyzing the captured event data into structural information immediately to generate a key event information set;
Based on the key event information set, carrying out event data analysis and efficiency data updating, analyzing the event information through a Pandas library of Python, including extracting event types, time stamps and player IDs, merging the event data with the original efficiency analysis data through a merge according to the player IDs, and referring to the instantaneity of the event, so that the data of each player can reflect the latest game performance, and generating an event updated data set;
Based on the event updated data set, performing data envelope analysis and random front edge analysis on the updated data by adopting a PyDEA library in Python and a front package of R software, setting input parameters and output parameters of a model again, adjusting parameters and running analysis commands, comprehensively referring to instant performance in a match, reflecting the efficiency state of each athlete, and generating an updated efficiency analysis result;
Based on the updated efficiency analysis result, adopting a hypothesis testing method, analyzing consistency and fairness in a scoring process, determining reliability of a scoring standard by using confidence interval estimation, optimizing the scoring standard, and generating a corrected scoring standard;
based on the corrected scoring standard, evaluating the information quantity of each competition index by adopting an entropy calculation method in an information theory, analyzing and selecting an index with the greatest influence on a scoring result through information gain, and adjusting a scoring system according to the information quantity, so as to generate an optimized scoring system;
Based on the optimized scoring system, a gradient elevator algorithm is adopted, a model is established by utilizing historical data, future performance and growth potential of athletes are predicted, and performance prediction models are generated by cross verification of optimized model parameters;
Based on the performance prediction model, a neural symbol learning method is adopted, the action and strategy of football is expressed through symbol logic, the relation between symbols and athlete performances is analyzed by using a deep learning model, and a symbolized skill and strategy evaluation result is generated;
Based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, an integrated learning method is adopted to construct an evaluation framework, a score generation and development scheme is provided for athletes, and a comprehensive performance evaluation result and a development plan are generated.
2. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: the efficiency analysis results comprise comprehensive scores, efficiency ranks and key performance indexes of athletes, the updated efficiency analysis results comprise adjusted athlete efficiency scores, real-time game performance updates and key event response evaluations, the corrected scoring criteria comprise newly set scoring thresholds, adjusted skill weights and scoring dimensions, the optimized scoring system comprises optimized scoring indexes, newly added performance dimensions and removed redundancy indexes, the performance prediction model comprises predicted growth tracks, key capacity improvement areas and potential performance bottlenecks, the symbolized skill and strategy evaluation results comprise symbolized skill mastery levels, strategy execution effects and personal improvement schemes, and the comprehensive performance evaluation results and development plans comprise personal comprehensive capacity evaluations, targeted development schemes and formulated training targets.
3. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: based on the updated efficiency analysis result, adopting a hypothesis testing method, analyzing consistency and fairness in the scoring process, determining reliability of the scoring standard by using confidence interval estimation, optimizing the scoring standard, generating the corrected scoring standard by the following specific steps,
Based on the updated efficiency analysis result, executing ttest _ind function through SciPy library of Python, setting significance level alpha as 0.05 for two groups of athlete efficiency score data sets through specified parameters, and comparing mean value differences of two independent samples to generate statistical results of hypothesis test;
based on the statistical result of the hypothesis test, using a norm/interval function of SciPy library, wherein parameters comprise a confidence level of 95% and a mean value and a standard deviation calculated based on sample data, providing a range estimation for the efficiency score of the athlete, and generating a confidence interval estimation result of the efficiency score;
and carrying out optimization operation of the scoring standard based on the statistical result of the hypothesis test and the confidence interval estimation result of the efficiency score, and generating a corrected scoring standard by adjusting the scoring parameter including modifying the weight and the threshold value in the scoring standard.
4. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: based on the corrected scoring standard, the information quantity of each competition index is evaluated by adopting an entropy calculation method in an information theory, the index with the largest influence on the scoring result is selected through information gain analysis, the scoring system is adjusted according to the index, the specific steps for generating the optimized scoring system are as follows,
Based on the corrected scoring standard, performing entropy calculation on the competition indexes by using a feature_selection module of SciKit-Learn library through Python, evaluating the information quantity of each index, identifying the indexes playing a key role in distinguishing the player performance, and generating an entropy calculation result of the competition indexes;
Based on the entropy calculation result of the competition index, an information gain analysis method is adopted, the entropy difference before and after the calculation index is removed, and the information gain analysis result is generated by selecting the index with the largest influence on the scoring result and identifying the most information value in the scoring system;
And adjusting a scoring system based on the information gain analysis result, and updating scoring system configuration, including optimizing index weights and adjusting scoring dimensions, by referring to the contribution of the indexes with the information gain to the accuracy of the scoring system, so that the scoring system fully reveals the real capability level of the athlete, and an optimized scoring system is generated.
5. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: based on the optimized scoring system, a gradient elevator algorithm is adopted, a model is established by utilizing historical data, the future performance and growth potential of athletes are predicted, the parameters of the optimized model are verified in a crossing way, the specific steps for generating a performance prediction model are as follows,
Based on the optimized scoring system, a gradient hoisting algorithm is selected for model initialization, gradientBoostingRegressor types are called through a scikit-learn library of Python, when a model is initialized, the n_ estimators parameter is set to be 100, the number of trees is defined, the learning_rate parameter is set to be 0.1, the learning rate is set, the max_depth parameter is set to be 4, the maximum depth of the tree is selected, and a prediction model basic structure is generated;
Based on the prediction model basic structure, adopting a cross_val_score function in scikit-learn library, setting cv parameters to be 5 for 5-fold cross validation, setting scoring parameters to be R2, adopting R2 as a scoring criterion, evaluating the performance of the model on multiple subsets, capturing the average performance index under the configuration of the model parameters, and generating a cross validation performance evaluation result;
Based on the cross-validation performance evaluation result, model parameters are adjusted and optimized, GRIDSEARCHCV is used for grid search of the parameters, multiple value ranges are set for the n_ estimators, learning _rate and the max_depth parameters for analysis, and the optimal model parameter combination is captured to generate a performance prediction model.
6. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: based on the performance prediction model, the action and strategy of football is expressed by symbol logic by adopting a neural symbol learning method, the relation between the symbol and the athlete performance is analyzed by utilizing a deep learning model, the specific steps for generating the symbolized skill and strategy evaluation result are as follows,
Based on the performance prediction model, constructing a neural symbol learning network, defining a network structure by using TensorFlow and Keras libraries, including setting an input layer to receive symbolized motion and strategy data, adding a plurality of hidden layers and adopting a ReLU as an activation function, and using a softmax function to perform multi-category prediction by an output layer to generate a deep learning network structure;
Based on the deep learning network structure, setting batch_size as 32 and epochs as 100, training a network by using a fit method, optimizing network weight, and generating a trained deep learning model;
Based on the trained deep learning model, a predict method is utilized to predict a test set, the output of the deep learning model is converted into scores for the skills and strategies of the athlete, skill and strategy feedback is provided for the athlete, and symbolic skill and strategy evaluation results are generated.
7. The neural network-based method for generating a college entrance examination football score according to claim 1, wherein: based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, an integrated learning method is adopted to construct an evaluation framework, a score generation and development scheme is provided for athletes, the specific steps for generating the comprehensive performance evaluation result and development scheme are as follows,
Based on the efficiency analysis result, the updated efficiency analysis result, the corrected scoring standard, the optimized scoring system, the performance prediction model, the symbolized skills and the strategy evaluation result, adopting an integrated learning method, integrating the prediction results of a plurality of models by utilizing VotingClassifier types in scikit-learn libraries, setting the weight distribution of a plurality of models, providing an evaluation and development plan for athletes, and generating a comprehensive evaluation framework;
Based on the comprehensive evaluation frame, refining and optimizing model configuration, and adjusting the weight of a single model in the integrated model according to a prediction result and current feedback to generate an optimized comprehensive evaluation frame;
Based on the optimized comprehensive evaluation framework, an integrated learning method is applied, and the comprehensive performance evaluation result and development plan are generated by finely analyzing the performance data of athletes in multiple dimensions and comprehensively reading and predicting the data.
8. The neural network-based college entrance examination football score generation system is characterized in that the neural network-based college entrance examination football score generation method is executed according to any one of claims 1-7, and comprises a data preprocessing module, an efficiency analysis module, an event processing module, a hypothesis testing module, a scoring system optimizing module, a performance predicting module, a skill strategy evaluating module and a comprehensive evaluation building module;
The data preprocessing module adopts a data cleaning algorithm based on the basic data and the competition performance data of athletes, uses Pandas library of Python to execute dropna and fillna functions to remove missing values and fill up, applies STANDARDSCALER functions to perform data standardization processing, and generates a preprocessed data set;
The efficiency analysis module executes efficiency analysis by using PyDEA libraries of Python based on the preprocessed data set by adopting a data envelope analysis algorithm, sets training time and physical quality as input parameters, uses shooting accuracy and running speed as output parameters, calls solve functions to calculate the efficiency value of each athlete, and generates a DEA efficiency analysis result;
The event processing module monitors real-time data flow by adopting a complex event processing algorithm based on a DEA efficiency analysis result and through an Esper technology of a Java platform, defines event pattern matching logic to identify a key event, analyzes captured event data into structured information and generates a key event information set;
The hypothesis testing module executes ttest _ind function through SciPy library of Python based on key event information set by adopting hypothesis testing algorithm, compares average difference of player efficiency scores, evaluates consistency and fairness of scoring process, and applies norm.interval function to carry out confidence interval estimation to generate corrected scoring standard;
The scoring system optimization module adopts entropy calculation and information gain analysis methods in the information theory based on the corrected scoring standard, evaluates and selects a competition index with the greatest influence on the scoring result by using a feature_selection module of SciKit-Learn library through Python, adjusts the scoring system according to the competition index, and generates an optimized scoring system;
The performance prediction module is used for constructing a prediction model through a scikit-learn library of Python based on an optimized scoring system by adopting a gradient hoisting algorithm, optimizing model parameters by using a cross verification method, predicting the future performance and growth potential of athletes, and generating a performance prediction model;
the skill strategy evaluation module adopts a neural symbol learning method based on a performance prediction model, uses TensorFlow and Keras libraries to construct a deep learning network, analyzes the relation between symbolized motion actions and strategies and player performances, provides evaluation for player skills and strategies, and generates symbolized skill and strategy evaluation results;
The comprehensive evaluation construction module integrates the prediction results of a plurality of models through scikit-learn library of Python by adopting an integrated learning method based on DEA efficiency analysis results, key event information sets, corrected scoring standards, optimized scoring systems, performance prediction models, symbolized skills and strategy evaluation results, provides score generation and development schemes for athletes, and generates comprehensive performance evaluation results and development plans.
CN202410579621.3A 2024-05-11 2024-05-11 Method and system for generating college entrance examination football score based on neural network Active CN118228889B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410579621.3A CN118228889B (en) 2024-05-11 2024-05-11 Method and system for generating college entrance examination football score based on neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410579621.3A CN118228889B (en) 2024-05-11 2024-05-11 Method and system for generating college entrance examination football score based on neural network

Publications (2)

Publication Number Publication Date
CN118228889A CN118228889A (en) 2024-06-21
CN118228889B true CN118228889B (en) 2024-08-20

Family

ID=91501076

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410579621.3A Active CN118228889B (en) 2024-05-11 2024-05-11 Method and system for generating college entrance examination football score based on neural network

Country Status (1)

Country Link
CN (1) CN118228889B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858719A (en) * 2017-11-30 2019-06-07 中国农业大学 Vineyard energy efficiency calculation method and device based on DEA and SFA
CN117972783A (en) * 2024-02-02 2024-05-03 厦门理工学院 Big data privacy protection method and system based on federal learning

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112766546B (en) * 2021-01-05 2022-04-26 上海大学 Machine learning-based football match win and loss real-time prediction method
WO2023283716A1 (en) * 2021-07-14 2023-01-19 Sportlogiq Inc. Systems and methods for detection, prediction, and value estimation of activities and events
CN116842070A (en) * 2023-06-09 2023-10-03 吉林亚泰中科医疗器械工程技术研究院股份有限公司 Football training monitoring and evaluating system based on big data analysis

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858719A (en) * 2017-11-30 2019-06-07 中国农业大学 Vineyard energy efficiency calculation method and device based on DEA and SFA
CN117972783A (en) * 2024-02-02 2024-05-03 厦门理工学院 Big data privacy protection method and system based on federal learning

Also Published As

Publication number Publication date
CN118228889A (en) 2024-06-21

Similar Documents

Publication Publication Date Title
Boroujeni et al. Discovery and temporal analysis of MOOC study patterns
CN109242149A (en) A kind of student performance early warning method and system excavated based on educational data
US11568753B2 (en) Systems and methods for learner growth tracking and assessments
CN109711707B (en) Comprehensive state evaluation method for ship power device
CN118135667B (en) Basketball testing method and basketball testing system based on AI
CN117390499A (en) Be applied to multiple sample detecting system that food pesticide remained and detected
Akçapınar et al. Modeling students’ academic performance based on their interactions in an online learning environment
CN118036475B (en) Optimal design method and system for carbon dioxide geological sequestration parameters
CN115409257A (en) Score distribution prediction method and system based on condition density estimation model
CN114358135B (en) MOOC (motion on-the-fly) conjugate prediction method realized by utilizing data enhancement and feature weighting
CN118228889B (en) Method and system for generating college entrance examination football score based on neural network
Azimi et al. Improving Students Performance in Small-Scale Online Courses--A Machine Learning-Based Intervention
KR102104007B1 (en) Apparatus and method for predicting result of game using predictive model of game result
CN110780660A (en) Tobacco production industry control system fault diagnosis method based on production state
CN118428827B (en) Teaching quality control method and system for modularized educational robot
CN117556381B (en) Knowledge level depth mining method and system for cross-disciplinary subjective test questions
Pavlik Jr et al. Modeling the EdNet dataset with logistic regression
CN118247108B (en) Regional earthquake security assessment method and system based on database technology
CN116739387B (en) Method and device for multidimensional analysis of data and computer storage medium
CN112138396B (en) Unmanned system simulation countermeasure-oriented intelligent body training method and system
Rabiha et al. Implementation of K-Nearest neighbor for classification teacher engagement profiling and interventions
Buddhika et al. Predictive Modeling for Enhancing MOOC Completion Rates: A Case Study
Olivares-Rodríguez et al. An actionable learning path-based model to predict and describe academic dropout
CN114676246A (en) Data processing method based on cloud computing
Li et al. Big data technology for teaching quality monitoring and improvement in higher education-joint K-means clustering algorithm and Apriori algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant