CN117172977A - Training suggestion generation method and system for trainee training - Google Patents
Training suggestion generation method and system for trainee training Download PDFInfo
- Publication number
- CN117172977A CN117172977A CN202311110218.8A CN202311110218A CN117172977A CN 117172977 A CN117172977 A CN 117172977A CN 202311110218 A CN202311110218 A CN 202311110218A CN 117172977 A CN117172977 A CN 117172977A
- Authority
- CN
- China
- Prior art keywords
- student
- training
- index
- data
- achievements
- 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.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 264
- 238000000034 method Methods 0.000 title claims abstract description 45
- 230000036541 health Effects 0.000 claims abstract description 85
- 238000007781 pre-processing Methods 0.000 claims abstract description 9
- 238000012216 screening Methods 0.000 claims abstract description 9
- 238000003066 decision tree Methods 0.000 claims description 33
- 238000011156 evaluation Methods 0.000 claims description 25
- 238000007637 random forest analysis Methods 0.000 claims description 24
- 238000004422 calculation algorithm Methods 0.000 claims description 18
- 238000009499 grossing Methods 0.000 claims description 18
- 238000010276 construction Methods 0.000 claims description 15
- 230000035945 sensitivity Effects 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 11
- 210000000577 adipose tissue Anatomy 0.000 claims description 8
- 230000009084 cardiovascular function Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 230000001419 dependent effect Effects 0.000 claims description 6
- 230000003862 health status Effects 0.000 claims description 2
- 230000005484 gravity Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 12
- 230000008901 benefit Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010219 correlation analysis Methods 0.000 description 3
- 230000000386 athletic effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000000611 regression analysis Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000000265 homogenisation Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000007786 learning performance Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses a training suggestion generation method and a training suggestion generation system for training a student, wherein the method comprises the following steps: obtaining a student data sample, carrying out data preprocessing on the student data sample, and constructing a student physical health condition index through an entropy method model; and calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training score based on a Pearson correlation coefficient method, screening to obtain characteristic indexes influencing the student score based on a set correlation coefficient threshold value, ranking and grading, carrying out multidimensional attribution on the score based on the graded grade, identifying the score characteristic type of the student, associating training schemes with similar movement requirements in a movement training expert database, and generating training advice of the student. The invention realizes the generation of the personalized training advice of the students, provides professional training guidance service for the students, improves the training effect of the students, and can adjust the training direction in time.
Description
Technical Field
The invention relates to the technical field of training data analysis, in particular to a training suggestion generation method and system for training a student.
Background
In the prior art, most of the training suggestions are formulated based on data, and most of the core algorithms adopt regression analysis, rule reasoning and the like. The key of the training suggestion generation is to acquire proper sports items and reasonable sports parameters, and the data-based training suggestion generation technology needs to establish a mapping relation between sports requirements and sports items for the selection of sports items. Next, a student's score is also generally predicted by training a score prediction model based on an algorithm such as regression analysis, and analyzing the historical score to predict the student's score performance at the next stage.
The above prior art solution, while addressing the basic need of training advice, still has some drawbacks, as follows:
1. in the prior art, the influence of a single index or factor on the later training direction is mostly considered, but the influence of multidimensional indexes such as physique and achievement influence factors of students are not fully considered.
2. Training advice in the prior art lacks the intelligent ability of discernment different group's demands, can't provide the guidance that pertinence is stronger for different students.
3. In the prior art, the automatic generation function of the training advice is generally realized, the achievement prediction model is not combined, the achievement of the learner is tracked and predicted, and the training advice is guided to be adjusted in time.
Therefore, how to intelligently identify the training requirements of different students according to multidimensional indexes such as the score attribution of the students, and the like, generate training suggestions of the students in a targeted manner, track and predict the score of the students, and timely adjust the training direction of the students is a problem to be solved.
Disclosure of Invention
In order to overcome the defects and shortcomings of the prior art, the invention provides a training suggestion generation method for training a student, aiming at the related requirements of the health and training effect of the student, the invention fuses big data application technology, combines historical performance data and expert library information by constructing a training suggestion association matching algorithm, a random forest and other prediction models, realizes the performance prediction and personalized training suggestion generation of the student, carries out professional exercise training guidance service for the student, improves the training effect of the student, predicts the later performance of the student according to the daily training performance of the student, monitors the training effect of the student and timely adjusts the training direction.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
the invention provides a training suggestion generation method aiming at trainee training, which comprises the following steps:
Obtaining a student data sample and a training suggestion expert database, wherein the student data sample comprises student information, student training results, student physical health data and student movement quality index data;
smoothing the history data of the students to obtain smooth values, and replacing abnormal achievements of the students or repairing missing data based on the smooth values;
constructing a comprehensive evaluation index system of the physical health condition of the student based on the physical health data of the student and the training score of the student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a sports ability dimension;
acquiring various index data of a comprehensive evaluation index system, and converting the absolute value of an index into a relative value;
acquiring the number of student data samples and the number of index data;
calculating the proportion of each index data in the index data of the trainee data sample, and calculating to obtain the entropy value of each index data;
calculating information entropy redundancy of each index dimension based on the entropy value of each index data, and calculating the weight of each index data based on the information entropy redundancy;
calculating to obtain the score of the physical health condition of the student based on the proportion of each index data in the index data of the student data sample and the weight of each index data;
The score of the physical health condition of the student is mapped, and the physical health condition of the student is estimated based on the mapped score, so that the physical health condition index of the student is obtained;
calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training results based on a Pearson correlation coefficient method by taking the student training results as dependent variables, and screening to obtain a characteristic index influencing the student results based on a set correlation coefficient threshold;
ranking and grading the numerical values corresponding to the characteristic indexes affecting the achievements of the students, and carrying out multidimensional attribution on the achievements based on the graded grades to obtain factor characteristics affecting the current achievements of the students, including the attribute of the quality of the exercises of the students and the attribute of the physical health of the students;
and identifying the score feature type of the student according to the factor feature affecting the current score of the student, analyzing the training requirements of the student according to the score feature type of the student, associating training schemes with similar sports requirements in a sports training expert database, and generating training suggestions of the student.
As a preferable technical scheme, the method further comprises a step of constructing a student achievement prediction model, and specifically comprises the following steps:
Obtaining training achievements of each subject of the student, calculating the median of the training achievements of each subject of the student, smoothing outliers and outliers of the achievements based on the median of the training achievements of each subject of the student, defining the usual achievements and the next-stage achievements of the student based on the training achievements of each subject of the student, and constructing hysteresis characteristic indexes of the usual achievements;
dividing the training results of each subject of the student after the smoothing treatment into a training data set and a testing data set;
and constructing a random forest model based on hysteresis characteristic indexes of the next-stage achievements and the usual achievements, wherein the random forest consists of a plurality of decision trees, training each decision tree by using a training data set and characteristics selected randomly, and carrying out an average or weighted average on the prediction results of the plurality of decision trees by using a recursive splitting algorithm in the training process of the decision tree to obtain a final student achievement prediction value, and adjusting a training scheme of a student based on the student achievement prediction value.
As a preferred technical scheme, the student information includes the name, grade and specialty of the student, the student physical health data includes the student's height, weight, body fat rate, body mass index, exercise intensity index, heart rate and cardiovascular function index, and the student exercise quality index data includes the student's speed, strength, endurance, flexibility and sensitivity indexes.
As a preferable technical scheme, the physical form dimension evaluation index system comprises the height, weight, body fat rate and body mass index of a student, the physical function dimension evaluation index system comprises the exercise intensity index, heart rate and cardiovascular function index of the student, and the exercise ability dimension evaluation index system comprises the achievements of each training subject of the student.
As an preferable technical solution, the step of obtaining each item of index data of the comprehensive evaluation index system and converting an absolute value of an index into a relative value includes:
dividing each index into a positive index and a negative index, and carrying out standardization processing on the positive index and the negative index, wherein the standardization processing is expressed as follows:
forward index:
negative index:
wherein x is ij Refers to the value, x of the ith record under the jth index ij ' represents x ij Normalized value, min (x j ) Refers to the minimum value, max (x j ) Refers to the maximum value under the j index.
As an preferable technical solution, the score of the physical health condition of the student is calculated based on the proportion of each index data in the index data of the student data sample and the weight of each index, which is specifically expressed as follows:
d j ＝1-E j ，j＝1，2，…，m
Wherein n represents the number of trainee data samples, m represents the number of index data in the trainee data samples, and p ij Representing the proportion of the jth index data in the index data of the ith trainee data sample, E j Entropy value d representing jth index data j Information entropy redundancy, ω, representing jth index data j The weight s of the j-th index data i A score representing the physical health status of the student of the ith student data sample.
As an optimal technical scheme, the characteristic indexes influencing the achievement of the students are obtained through screening based on the set correlation coefficient threshold, wherein the characteristic indexes influencing the achievement of the students comprise physical health condition indexes, speed indexes, strength indexes, endurance indexes, sensitivity indexes and flexibility indexes of the students;
the student movement quality attribution comprises a speed index, a force index quantity, a endurance index, a flexibility index and a sensitivity index attribution;
the attribution of the physical health of the students comprises attribution of physical health condition indexes of the students.
As an preferable technical scheme, the ranking and grading are performed on the values corresponding to the characteristic indexes affecting the achievement of the learner, specifically, the ranking is performed on the values corresponding to the characteristic indexes affecting the achievement of the learner according to pareto rule, and the grading is performed on the values corresponding to the characteristic indexes after ranking according to the second-eighth rule.
The invention also provides a training suggestion generation system for training a student, which comprises the following steps: the system comprises a data acquisition module, a data preprocessing module, a student physical health condition index construction module, a correlation characteristic index construction module, a multidimensional attribution module and a training suggestion generation module;
the data acquisition module is used for acquiring a student data sample and a training suggestion expert database, wherein the student data sample comprises student information, student training results, student physical health data and student movement quality index data;
the data preprocessing module is used for carrying out smoothing processing on the history data of the students to obtain smooth values, and replacing abnormal achievements of the students or repairing missing data based on the smooth values;
the student physical health condition index construction module is used for constructing student physical health condition indexes and specifically comprises the following steps:
constructing a comprehensive evaluation index system of the physical health condition of the student based on the physical health data of the student and the training score of the student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a sports ability dimension;
acquiring various index data of a comprehensive evaluation index system, and converting the absolute value of an index into a relative value;
Acquiring the number of student data samples and the number of index data;
calculating the proportion of each index data in the index data of the trainee data sample, and calculating to obtain the entropy value of each index data;
calculating information entropy redundancy of each index dimension based on the entropy value of each index data, and calculating the weight of each index data based on the information entropy redundancy;
calculating to obtain the score of the physical health condition of the student based on the proportion of each index data in the index data of the student data sample and the weight of each index data;
the score of the physical health condition of the student is mapped, and the physical health condition of the student is estimated based on the mapped score, so that the physical health condition index of the student is obtained;
the relevance characteristic index construction module is used for constructing characteristic indexes affecting the achievement of students, and specifically comprises the following steps:
calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training results based on a Pearson correlation coefficient method by taking the student training results as dependent variables, and screening to obtain a characteristic index influencing the student results based on a set correlation coefficient threshold;
The multi-dimensional attribution module is used for ranking and grading the numerical values corresponding to the characteristic indexes affecting the achievement of the students, and carrying out multi-dimensional attribution on the achievement based on the graded grades, wherein the multi-dimensional attribution module comprises the attribution of the quality of the exercises of the students and the attribution of the physical health of the students;
the training suggestion generation module is used for identifying the score feature type of the student according to the factor feature affecting the current score of the student, analyzing the training requirement of the student according to the score feature type of the student, associating training schemes with similar motion requirements in the motion training expert library, and generating training suggestions of the student.
As a preferred technical solution, the system further includes a student score prediction model construction module, where the student score prediction model construction module is configured to construct a student score prediction model, and specifically includes:
obtaining training achievements of each subject of the student, calculating the median of the training achievements of each subject of the student, smoothing outliers and outliers of the achievements based on the median of the training achievements of each subject of the student, defining the usual achievements and the next-stage achievements of the student based on the training achievements of each subject of the student, and constructing hysteresis characteristic indexes of the usual achievements;
dividing the training results of each subject of the student after the smoothing treatment into a training data set and a testing data set;
And constructing a random forest model based on hysteresis characteristic indexes of the next-stage achievements and the usual achievements, wherein the random forest consists of a plurality of decision trees, training each decision tree by using a training data set and characteristics selected randomly, and carrying out an average or weighted average on the prediction results of the plurality of decision trees by using a recursive splitting algorithm in the training process of the decision tree to obtain a final student achievement prediction value, and adjusting a training scheme of a student based on the student achievement prediction value.
Compared with the prior art, the invention has the following advantages and beneficial effects:
(1) In the prior art, only the influence of a single index or factor on the later training direction is considered, the influence of multi-dimensional indexes such as physique, historical performance and the like of a student on the later training direction is fully considered, a training suggestion association matching algorithm is constructed, the training requirements of the student in the current stage are intelligently identified, more targeted guidance is provided for different students, and individualized training suggestions of the student are generated by combining expert database information, so that multi-factor comprehensive analysis is realized.
(2) According to the training recommendation method, the training recommendation correlation matching algorithm and the random forest prediction model are constructed, the personal score trend prediction algorithm model is fused on the basis of personalized training recommendation, future scores of students are tracked and predicted, timely adjustment of training directions of the students is assisted, and training quality and training effect of the students are improved.
Drawings
FIG. 1 is a flow chart of a training advice generation method for trainee training according to the present invention.
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.
Example 1
The embodiment provides a training suggestion generation method for training a student, which is used for making an intelligent matching strategy based on basic information of the student, daily training results, a training suggestion expert database, bracelet monitoring and other data and automatically generating training suggestions of the student by using a recommendation algorithm such as association matching; meanwhile, a random forest prediction model is constructed to track and monitor the score change trend of the trainees, track and predict future scores of the trainees, assist in timely adjustment of training directions of the trainees, and improve training quality and effect of the trainees;
when intelligent training advice is realized, an intelligent training advice association matching algorithm is constructed based on a training advice expert library of the trainees, the influence of multi-dimensional indexes such as comprehensive physique, achievement attribution and the like of the trainees on the later training direction is fully considered by an algorithm model, the training requirements of different trainees are intelligently identified, training advice of the trainees is generated in a targeted manner, and guidance is provided for the later training of the trainees;
When the achievement prediction is realized, the historical achievement of the student is analyzed, the relation between the achievement of each subject and the achievement of the next stage in daily training is explored, a random forest prediction model is constructed, the achievement prediction of the student is realized, the achievement of the student is tracked and predicted, the training direction of the student is timely adjusted, and the training quality is improved.
The method comprises the following specific steps:
s1: data preparation, in this embodiment, obtaining data of the height, weight, grade, specialty, body fat rate, body mass index, exercise intensity index, heart rate, cardiovascular function and the purpose of each training department of the student, speed, strength, endurance, flexibility and sensitivity as input data of a performance attribution model;
the physical health condition index of the student is obtained by constructing an entropy method model for comprehensive evaluation based on the height, weight, body fat rate, body mass index, exercise intensity index, heart rate, cardiovascular function of the student and score index of each training subject of the student; the speed, strength, endurance, flexibility and sensitivity indexes are measured by a sports ability testing instrument;
s2: and (5) preprocessing data. The history data of the trainee is used for smoothing, and the median of the recent week of the trainee is used as a smoothing value to replace and repair abnormal score or missing data of the trainee.
S3: characteristic engineering: the method comprehensively evaluates the physical health condition index of the student based on the entropy method model, and specifically comprises the following steps:
s31: constructing a comprehensive evaluation index system of the physical health condition of a student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a movement capability dimension, and the physical form dimension considers factors such as the height, the weight, the body fat rate, the body mass index and the like of the student; the physical function dimension considers the exercise intensity index, heart rate and cardiovascular function factors of the trainee; the athletic ability dimension takes into account the performance of each training subject of the learner.
S32: normalization processing of index data: because the measurement units of the indexes are not uniform, before the comprehensive indexes are calculated by the measurement units, the standardization treatment is carried out, namely, the absolute values of the indexes are converted into relative values, so that the homogenization problem of various different index values is solved.
In this embodiment, each index is divided into a positive index and a negative index, where the positive index is also called a benefit index, and represents an index that progresses upward or forward, and increases, and the larger the index value, the better the smaller the index value, the worse the index value; the negative index has the opposite effect to the positive index, the smaller the value is, the better or the larger the value is, the less good the value is, and the specific calculation formula is expressed as follows:
Forward index:
negative index:
wherein x is ij Refers to the value, x of the ith record under the jth index ij ' represents x ij Normalized value, min (x j ) Refers to the minimum value, max (x j ) Refers to the maximum value under the j index.
S33: obtaining the number n of rows and the number m of columns of data, wherein n represents the number of data samples of a student, m represents the number of indexes, and the data samples of the student of the entropy method model consist of the height, weight, body fat rate, body mass index of the student, exercise intensity index of the student, heart rate, cardiovascular function and the score of each training subject;
s34: calculating the proportion of the ith record under the jth index:
s35: calculating the entropy value of the j-th index:
wherein k represents a constant, and the calculation formula is: k=1/ln (n);
s36: the information entropy redundancy (difference) of each index dimension is calculated as follows:
d j ＝1-E j ，j＝1，2，…，m
due to the information entropy E j Is used for measuring the information utility value of j indexes, and when the index is completely unordered, E j At this time, the utility value of the index data to the comprehensive evaluation is zero, so a difference is needed to calculate the information utility value of a certain index;
s37: the weights of the indexes are calculated, and the calculation formula is as follows:
S38: based on characteristic data of the students, according to weights of various indexes of the students, and combining actual conditions of the indexes, evaluating physical health conditions of the students;
s39: based on the score of the physical health condition of the student obtained in S38, the score mapping is evaluated, and in this embodiment, the score is mapped to 60-100, and when the score is higher, the better the physical health condition of the student is indicated.
S4: and (5) factor correlation analysis.
Constructing a factor correlation analysis model, and screening relevant factors of achievements from various exercise attribution influence indexes such as height, weight, grade, specialty, physical health condition, speed, strength, endurance, flexibility, sensitivity and the like of a student to serve as model input features;
the factor correlation analysis model selects a pearson correlation coefficient method, takes the achievement of a student as a dependent variable Y, takes the height, weight, grade, specialty, physical health condition, speed, strength, endurance, flexibility and sensitivity as independent variables X, calculates the correlation coefficient of each X and Y, and defines the correlation coefficient r as the product of covariance of two variables X and Y divided by standard deviation of the two variables, and has the following calculation formula:
the coefficient r has the following properties:
when r is more than or equal to-1 and less than or equal to 1, the larger the absolute value of r is, the stronger the correlation degree between two variables is shown;
When r is more than 0 and less than or equal to 1, indicating that positive correlation exists between the two variables;
when r=1, then it is indicated that there is a complete positive correlation between the variables;
when r is less than or equal to 0 and less than-1, indicating that negative correlation exists between the two variables;
when r= -1, it indicates that there is a completely negative correlation between the variables;
when r=0, a radio correlation between the two variables is indicated;
setting a correlation coefficient threshold, filtering characteristic indexes with too low correlation coefficients with the achievement of the students, and reserving the characteristic indexes which are larger than or equal to the correlation coefficient threshold, wherein the current correlation coefficient threshold is preferably 0.6.
The finally screened characteristic indexes affecting the achievement of the students are physical health conditions, speed indexes, strength indexes, endurance indexes, sensitivity indexes and flexibility indexes of the students.
S5: constructing a score attribution model;
s51: obtaining result data of a factor-related analysis model, and extracting relevant factors of achievements such as physical health conditions, speed indexes, strength indexes, endurance indexes, sensitivity indexes, flexibility indexes and the like of a student as model characteristics;
s52: according to the pareto rule, the grades of all factors of the students are calculated, the grades of which elements are poor and influence the achievement of the students, firstly, the values of the physical health condition, the speed index, the strength index, the endurance index, the sensitivity index and the flexibility index of the students are ranked respectively, and then the grades of the above-mentioned indexes of the students are classified into four grades of excellent, good, medium and bad according to the second eight rule;
S53: the multi-dimensional attribution of the achievement mainly comprises attribution of sports quality (speed, strength, endurance, flexibility, sensitivity and the like) and attribution of physical health (physical health condition), and the index with the grade of poor is summarized and is a main factor causing poor achievement of the student.
In the embodiment, relevant factors influencing the training performance of the trainee are quantified, performance influencing factors are extracted, and the technical means such as a machine learning algorithm, a data mining algorithm and the like are utilized to carry out deep analysis on attribute information and behavior information of the trainee, so that main factors influencing the performance of the trainee are intelligently positioned, and training performance attribution is realized;
the training score attribution can provide important reference data for students, is beneficial to the students to comprehensively know the performance and characteristics in the learning process, and discovers the advantages and the disadvantages of the students, so that more accurate training plans and targets can be adjusted or formulated in a targeted manner, personalized culture of the students is promoted, and the learning efficiency and the learning quality of the students are improved; on the other hand, training score attribution can help the instructor to objectively and comprehensively understand the capability and level of the trainee, quickly position the reason of poor score of the trainee, pertinently design courses and teaching contents, promote the comprehensive development of the trainee and improve the teaching effect; and secondly, training score attribution can help the instructor to reasonably utilize the information of the learner, optimize the configuration of teaching resources and improve the teaching efficiency, and thirdly, the training score attribution can help the instructor to find problems existing in self teaching, adjust teaching strategies according to specific conditions and improve the teaching quality, and finally, the data obtained by the training score attribution can provide decision support for schools to help the schools to make education strategies and development plans.
S6: constructing a training suggestion model;
s61: obtaining attribution results of the student achievements, and obtaining factor characteristics affecting the current achievements of the students;
s62: identifying the performance characteristic type of the student according to the factor characteristics affecting the current performance of the student, namely whether the performance of the student is affected by the physical health condition of the student or is caused by poor sports quality indexes such as speed, strength, endurance, flexibility, sensitivity and the like;
according to the performance feature type of the student, the training requirement of the student is analyzed, namely whether human intervention is introduced preferentially to promote the physical health of the student or whether a certain sports quality index of the student is concentrated to be promoted is judged;
dividing the students into corresponding groups to which the training belongs through personalized analysis of the exercise requirement, such as dividing the physical health condition into the same group of groups, manually intervening the training of the students, dividing the achievement into the same group of groups mainly influenced by the speed index, and concentrating on the improvement of the speed direction in later training;
s63: by means of the association matching strategy and combining with training requirements of students, training schemes with similar motion requirements in a training expert library of sports training are automatically associated, corresponding sports training prescriptions are matched for the students, training suggestion schemes are generated, and training suggestions of the students are generated to guide training directions of the students in later stages;
If the requirement of the student is to enhance the physical health condition of the student, the training scheme which can be matched with the same motion requirement in the expert database according to the motion requirement of the student is to manually intervene the training of the student, the training is customized, when the comprehensive physical health condition of the student is poor, a student injury prevention mechanism is started, the later training of the student is recommended to manually intervene by the student, and the training is arranged according to the actual condition; if the exercise requirement is focused on the speed improvement index, the training proposal scheme given by the associated expert database is a detailed training project, training sequence and training time for helping the learner to improve the speed at the stage;
s7: the embodiment further includes a step of constructing a student score prediction model, predicting the score of the student at the next stage, monitoring the student to follow the score change after the training advice, that is, monitoring the effect of the training advice on the improvement of the score of the student, which is beneficial to timely finding whether the training scheme of the student is reasonable in the training process and making an adjustment, and the specific steps include:
s71: data preparation, namely adopting daily training results of each subject of a student as input data of a student result prediction model;
s72: and (3) feature engineering, preprocessing data, cleaning the score data of each subject of the student, and smoothing outliers and outliers of the results by using recent median of the results of the student.
S73: the achievement prediction model is constructed, and concretely comprises the following steps:
s731: defining a usual score and a next-stage score of the student based on the training score of the student;
s732: rolling back a window on the usual achievements of each subject of the student, and constructing a hysteresis characteristic index of the student achievements based on the median of the recent achievements of each subject of the student;
s733: based on the results of step S731 and step S732, training performance data of the trainee is divided into a training data set and a test data set.
S734: firstly, a prediction model is built, and a random forest model is built by combining the hysteresis characteristic indexes of the achievements of the lower stage of the students and the usual achievements based on the training set data in the step S733. Wherein the random forest is composed of a plurality of decision trees, each of which is constructed based on randomly selected training set data samples and features. Next, feature selection is performed. As the nodes of each decision tree split, the random forest will select features from a random subset. This helps reduce the correlation between features and improves the generalization ability of the model. The decision tree is then trained. For each decision tree, training is performed using randomly selected training set data samples and features. The training process of decision trees typically uses a recursive splitting algorithm to split according to the feature's unreliability. Finally, the random forest model averages or weights and averages the prediction results of the decision trees to obtain a final prediction value, namely, the student score prediction model is successfully constructed, and the random forest model can be used for mining a logic rule between the usual score and the score of the next stage of each subject.
The specific process of random forest prediction is as follows:
(1) Assuming that the original training dataset is denoted as N and that the multiple trees of the decision tree constructed by the random forest are denoted as M, then m=m 1 ，m 2 ，…x t 。
(2) The original data N is resampled by bootstrap (self-help method), namely N times of random extraction are repeated in N, and the N times of random extraction are used as a new training set (the training set of each decision tree has the same size as the original data set).
(2) And (6) constructing a decision tree. And constructing a decision tree for each training set, repeating the process for i times, and constructing i decision trees.
(3) By H (x) Represents a random forest model, then h 1 (X)h 2 (X)…h i (X) is a plurality of decision trees, X represents sample characteristic attributes, I represents an indication function, and a random forestThe regression prediction formula is as follows:
H (x) ＝avg(∑h i (x))
(4) And counting the weighted average of all the decision tree results to obtain a random forest regression prediction result.
S74: and evaluating a performance prediction model.
Calculating evaluation indexes of the achievement prediction, such as average absolute error, standard error, average absolute percentage error and R square value, by using the test data set in the step S733, and evaluating the accuracy of the achievement prediction of the students;
s75: and (5) applying a achievement prediction model.
And predicting the end of the period of the student by using the daily training results of each subject of the student and applying the result prediction model in the step S734, thereby monitoring the effect of the training advice on the improvement of the results of the student and timely adjusting the training scheme of the student.
In the embodiment, a training suggestion association matching algorithm and a random forest model are constructed to realize the performance prediction and personalized training suggestion generation of students, the training suggestions need to have the capability of intelligently identifying the requirements of different students, and the influence of multi-dimensional indexes such as physique conditions, various performance influence factors and the like of the students on the training suggestions and the performance prediction is considered, so that a training scheme suitable for the students can be provided for the students rapidly, the autonomous learning consciousness and the autonomous learning ability of the students are stimulated and cultivated, the students participate in the learning process more actively, and the learning efficiency and the learning quality are improved; secondly, the embodiment can help students to learn about own learning performance and progress, enhance self-confidence and achievement sense of the students, and improve enthusiasm of the students to participate in training; on the other hand, the method and the device can help the instructor to quickly know the academic difference of the instructor, realize the establishment of a personalized training scheme of the instructor, provide personalized chemistry Xi Jian protocol and instruction for each instructor, enable the instructor to better exert own advantages and overcome own defects; furthermore, the training effect of the learner can be monitored in time by the aid of the teaching staff, training conditions and requirements of the learner are better known, training strategies are timely adjusted according to the training effect, and teaching quality is improved.
Example 2
The present embodiment provides a training advice generation system for training of a learner, including: the system comprises a data acquisition module, a data preprocessing module, a student physical health condition index construction module, a correlation characteristic index construction module, a multidimensional attribution module and a training suggestion generation module;
in this embodiment, the data acquisition module is configured to acquire a trainee data sample and a training advice expert database, where the trainee data sample includes trainee information, trainee training results, trainee physical health data, and trainee athletic quality index data;
in this embodiment, the data preprocessing module is configured to perform smoothing processing on the history data of the learner to obtain a smoothed value, and replace abnormal score of the learner or repair missing data based on the smoothed value;
in this embodiment, the student physical health condition index construction module is configured to construct a student physical health condition index, and specifically includes:
constructing a comprehensive evaluation index system of the physical health condition of the student based on the physical health data of the student and the training score of the student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a sports ability dimension;
acquiring various index data of a comprehensive evaluation index system, and converting the absolute value of an index into a relative value;
Acquiring the number of student data samples and the number of index data;
calculating the proportion of each index data in the index data of the trainee data sample, and calculating to obtain the entropy value of each index data;
calculating information entropy redundancy of each index dimension based on the entropy value of each index data, and calculating the weight of each index data based on the information entropy redundancy;
calculating to obtain the score of the physical health condition of the student based on the proportion of each index data in the index data of the student data sample and the weight of each index data;
the score of the physical health condition of the student is mapped, and the physical health condition of the student is estimated based on the mapped score, so that the physical health condition index of the student is obtained;
in this embodiment, the relevance feature index building module is configured to build feature indexes that affect the achievement of the learner, and specifically includes:
calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training results based on a Pearson correlation coefficient method by taking the student training results as dependent variables, and screening to obtain a characteristic index influencing the student results based on a set correlation coefficient threshold;
In this embodiment, the multidimensional attribution module is configured to rank and rank the values corresponding to the feature indexes that affect the achievement of the learner, and perform multidimensional attribution on the achievement based on the ranked grades, including attribution of the quality of the movement of the learner, and attribution of the physical health of the learner;
in this embodiment, the training suggestion generating module is configured to identify a score feature type of a learner according to a factor feature affecting a current score of the learner, analyze a training requirement of the learner according to the score feature type of the learner, and associate training schemes with similar training requirements in the training expert database to generate training suggestions of the learner.
In this embodiment, the system further includes a student score prediction model building module, where the student score prediction model building module is configured to build a student score prediction model, and specifically includes:
obtaining training achievements of each subject of the student, calculating the median of the training achievements of each subject of the student, smoothing outliers and outliers of the achievements based on the median of the training achievements of each subject of the student, defining the usual achievements and the next-stage achievements of the student based on the training achievements of each subject of the student, and constructing hysteresis characteristic indexes of the usual achievements;
dividing the training results of each subject of the student after the smoothing treatment into a training data set and a testing data set;
And constructing a random forest model based on hysteresis characteristic indexes of the next-stage achievements and the usual achievements, wherein the random forest consists of a plurality of decision trees, training each decision tree by using a training data set and characteristics selected randomly, and carrying out an average or weighted average on the prediction results of the plurality of decision trees by using a recursive splitting algorithm in the training process of the decision tree to obtain a final student achievement prediction value, and adjusting a training scheme of a student based on the student achievement prediction value.
The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.
Claims (10)
1. A training advice generation method for trainee training, comprising the steps of:
obtaining a student data sample and a training suggestion expert database, wherein the student data sample comprises student information, student training results, student physical health data and student movement quality index data;
smoothing the history data of the students to obtain smooth values, and replacing abnormal achievements of the students or repairing missing data based on the smooth values;
Constructing a comprehensive evaluation index system of the physical health condition of the student based on the physical health data of the student and the training score of the student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a sports ability dimension;
acquiring various index data of a comprehensive evaluation index system, and converting the absolute value of an index into a relative value;
acquiring the number of student data samples and the number of index data;
calculating the proportion of each index data in the index data of the trainee data sample, and calculating to obtain the entropy value of each index data;
calculating information entropy redundancy of each index dimension based on the entropy value of each index data, and calculating the weight of each index data based on the information entropy redundancy;
calculating to obtain the score of the physical health condition of the student based on the proportion of each index data in the index data of the student data sample and the weight of each index data;
the score of the physical health condition of the student is mapped, and the physical health condition of the student is estimated based on the mapped score, so that the physical health condition index of the student is obtained;
calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training results based on a Pearson correlation coefficient method by taking the student training results as dependent variables, and screening to obtain a characteristic index influencing the student results based on a set correlation coefficient threshold;
Ranking and grading the numerical values corresponding to the characteristic indexes affecting the achievements of the students, and carrying out multidimensional attribution on the achievements based on the graded grades to obtain factor characteristics affecting the current achievements of the students, including the attribute of the quality of the exercises of the students and the attribute of the physical health of the students;
and identifying the score feature type of the student according to the factor feature affecting the current score of the student, analyzing the training requirements of the student according to the score feature type of the student, associating training schemes with similar sports requirements in a sports training expert database, and generating training suggestions of the student.
2. The training advice generation method for trainee training of claim 1, further comprising a trainee performance prediction model construction step, comprising:
obtaining training achievements of each subject of the student, calculating the median of the training achievements of each subject of the student, smoothing outliers and outliers of the achievements based on the median of the training achievements of each subject of the student, defining the usual achievements and the next-stage achievements of the student based on the training achievements of each subject of the student, and constructing hysteresis characteristic indexes of the usual achievements;
dividing the training results of each subject of the student after the smoothing treatment into a training data set and a testing data set;
And constructing a random forest model based on hysteresis characteristic indexes of the next-stage achievements and the usual achievements, wherein the random forest consists of a plurality of decision trees, training each decision tree by using a training data set and characteristics selected randomly, and carrying out an average or weighted average on the prediction results of the plurality of decision trees by using a recursive splitting algorithm in the training process of the decision tree to obtain a final student achievement prediction value, and adjusting a training scheme of a student based on the student achievement prediction value.
3. The training advice generation method for training of a learner of claim 1, wherein the learner information includes a name, a grade, and a specialty of the learner, the learner physical health data includes a height, a weight, a body fat rate, a body mass index, a exercise intensity index, a heart rate, and a cardiovascular function index of the learner, and the learner exercise quality index data includes speed, strength, endurance, flexibility, and sensitivity indexes of the learner.
4. The training advice generation method for training of a learner of claim 1, wherein the physical form dimension assessment index system comprises a height, a weight, a body fat rate, and a body mass index of the learner, the physical function dimension assessment index system comprises a exercise intensity index, a heart rate, and a cardiovascular function index of the learner, and the exercise ability dimension assessment index system comprises a performance of each training subject of the learner.
5. The training advice generation method for training of a trainee according to claim 1, wherein the step of obtaining each item of index data of the comprehensive evaluation index system and converting an absolute value of an index into a relative value comprises:
dividing each index into a positive index and a negative index, and carrying out standardization processing on the positive index and the negative index, wherein the standardization processing is expressed as follows:
forward index:
negative index:
wherein x is ij Refers to the value, x of the ith record under the jth index ij ' represents x ij Normalized value, min (x j ) Refers to the minimum value, max (x j ) Refers to the maximum value under the j index.
6. The training advice generation method for training of a learner according to claim 1, wherein the score of the physical health condition of the learner is calculated based on the specific gravity of each index data in the index data of the learner data sample and the weight of each index, specifically expressed as:
wherein n represents the number of trainee data samples, m represents the number of index data in the trainee data samples, and p ij Representing the proportion of the jth index data in the index data of the ith trainee data sample, E j Entropy value d representing jth index data j Information entropy redundancy, ω, representing jth index data j The weight s of the j-th index data i A score representing the physical health status of the student of the ith student data sample.
7. The training advice generation method for training of a learner according to claim 1, wherein the characteristic index influencing the performance of the learner is obtained by screening based on a set correlation coefficient threshold, and the characteristic index influencing the performance of the learner includes a physical health condition index, a speed index, a strength index, a endurance index, a sensitivity index, and a flexibility index of the learner;
the student movement quality attribution comprises a speed index, a force index quantity, a endurance index, a flexibility index and a sensitivity index attribution;
the attribution of the physical health of the students comprises attribution of physical health condition indexes of the students.
8. The training advice generation method for training of a learner according to claim 1, wherein the ranking and grading are performed on the values corresponding to the characteristic indexes affecting the performance of the learner, specifically, the ranking is performed on the values corresponding to the characteristic indexes affecting the performance of the learner according to pareto rule, and the grading is performed on the values corresponding to the characteristic indexes after ranking according to the second eight rule.
9. A training advice generation system for trainee training, comprising: the system comprises a data acquisition module, a data preprocessing module, a student physical health condition index construction module, a correlation characteristic index construction module, a multidimensional attribution module and a training suggestion generation module;
the data acquisition module is used for acquiring a student data sample and a training suggestion expert database, wherein the student data sample comprises student information, student training results, student physical health data and student movement quality index data;
the data preprocessing module is used for carrying out smoothing processing on the history data of the students to obtain smooth values, and replacing abnormal achievements of the students or repairing missing data based on the smooth values;
the student physical health condition index construction module is used for constructing student physical health condition indexes and specifically comprises the following steps:
constructing a comprehensive evaluation index system of the physical health condition of the student based on the physical health data of the student and the training score of the student, wherein the comprehensive evaluation index system comprises a physical form dimension, a physical function dimension and a sports ability dimension;
acquiring various index data of a comprehensive evaluation index system, and converting the absolute value of an index into a relative value;
Acquiring the number of student data samples and the number of index data;
calculating the proportion of each index data in the index data of the trainee data sample, and calculating to obtain the entropy value of each index data;
calculating information entropy redundancy of each index dimension based on the entropy value of each index data, and calculating the weight of each index data based on the information entropy redundancy;
calculating to obtain the score of the physical health condition of the student based on the proportion of each index data in the index data of the student data sample and the weight of each index data;
the score of the physical health condition of the student is mapped, and the physical health condition of the student is estimated based on the mapped score, so that the physical health condition index of the student is obtained;
the relevance characteristic index construction module is used for constructing characteristic indexes affecting the achievement of students, and specifically comprises the following steps:
calculating the correlation coefficient of each data of student information, student physical health condition indexes and student movement quality index data and student training results based on a Pearson correlation coefficient method by taking the student training results as dependent variables, and screening to obtain a characteristic index influencing the student results based on a set correlation coefficient threshold;
The multi-dimensional attribution module is used for ranking and grading the numerical values corresponding to the characteristic indexes affecting the achievement of the students, and carrying out multi-dimensional attribution on the achievement based on the graded grades, wherein the multi-dimensional attribution module comprises the attribution of the quality of the exercises of the students and the attribution of the physical health of the students;
the training suggestion generation module is used for identifying the score feature type of the student according to the factor feature affecting the current score of the student, analyzing the training requirement of the student according to the score feature type of the student, associating training schemes with similar motion requirements in the motion training expert library, and generating training suggestions of the student.
10. The training advice generation system for trainee training of claim 9, further comprising a trainee score prediction model construction module for constructing a trainee score prediction model, comprising:
obtaining training achievements of each subject of the student, calculating the median of the training achievements of each subject of the student, smoothing outliers and outliers of the achievements based on the median of the training achievements of each subject of the student, defining the usual achievements and the next-stage achievements of the student based on the training achievements of each subject of the student, and constructing hysteresis characteristic indexes of the usual achievements;
Dividing the training results of each subject of the student after the smoothing treatment into a training data set and a testing data set;
constructing a random forest model based on hysteresis characteristic indexes of the next-stage achievements and the usual achievements, wherein,
the random forest is composed of a plurality of decision trees, each decision tree is trained by using a training data set and features which are selected randomly, a recursive splitting algorithm is used in the training process of the decision tree, the prediction results of the decision trees are averaged or weighted averaged to obtain a final student score prediction value, and a training scheme of a student is adjusted based on the student score prediction value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311110218.8A CN117172977A (en) | 2023-08-30 | 2023-08-30 | Training suggestion generation method and system for trainee training |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311110218.8A CN117172977A (en) | 2023-08-30 | 2023-08-30 | Training suggestion generation method and system for trainee training |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117172977A true CN117172977A (en) | 2023-12-05 |
Family
ID=88932984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311110218.8A Pending CN117172977A (en) | 2023-08-30 | 2023-08-30 | Training suggestion generation method and system for trainee training |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117172977A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117711627A (en) * | 2024-02-06 | 2024-03-15 | 中国民用航空飞行学院 | Health risk prediction treatment method and system for civil aviation flight trainee in flight training process |
-
2023
- 2023-08-30 CN CN202311110218.8A patent/CN117172977A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117711627A (en) * | 2024-02-06 | 2024-03-15 | 中国民用航空飞行学院 | Health risk prediction treatment method and system for civil aviation flight trainee in flight training process |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109919810B (en) | Student modeling and personalized course recommendation method in online learning system | |
| CN109858797B (en) | Multi-dimensional informatics analysis method based on knowledge network accurate online education system | |
| CN111274411A (en) | Course recommendation method and device, electronic equipment and readable storage medium | |
| CN107274020B (en) | Learner subject total measured result prediction system and method based on collaborative filtering thought | |
| CN106779166A (en) | Grasp the forecasting system and method for state in a kind of knowledge point based on data-driven | |
| CN111914162B (en) | Method for guiding personalized learning scheme based on knowledge graph | |
| CN114822763A (en) | Personalized exercise prescription recommendation method driven by exercise data | |
| CN117172977A (en) | Training suggestion generation method and system for trainee training | |
| CN106614273A (en) | Pet feeding method and system based on big data analysis of Internet of Things | |
| CN111626372A (en) | Online teaching supervision management method and system | |
| US20230146132A1 (en) | Systems and methods for learner growth tracking and assessments | |
| Stanković et al. | Artificial neural network model for prediction of students’ success in learning programming | |
| CN117313947A (en) | Learning early warning method based on random forest and tuna group optimization algorithm | |
| CN117350905A (en) | Sports teaching management system based on big data | |
| CN117352114A (en) | Recommendation method and system of psychological assessment scale based on clustering algorithm | |
| CN117291770A (en) | Visual user data assessment method and system for teaching effect | |
| JP7439914B2 (en) | Prediction device, prediction method, and program | |
| CN112883284A (en) | Online learning system based on network and data analysis and test question recommendation method | |
| CN116842070A (en) | Football training monitoring and evaluating system based on big data analysis | |
| CN117540108B (en) | Intelligent recommendation answering system based on examination point data distributed summary | |
| Li | An evaluation of college students' involvement in English learning behaviour based on fuzzy comprehensive evaluation | |
| Xu et al. | A Study and Analysis of Predicting College Students’ Final Exam Scores by Integrating Physical Fitness Test Data and Poverty Level Information | |
| Tang | Research on Stable Prediction Method of Online Learning Effect Based on Decision Tree | |
| Babaee Khobdeh et al. | A novel method for clustering basketball players with data mining and hierarchical algorithm | |
| Khobdeh et al. | hierarchical 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 |