The Establishment Of Predict Ion Model Of Children’s Breaststroke Learning Outcomes Based On Physique Variables And The Feature, Effects Analysis Of Children’s Sensory Integration Levels Between Different Breaststroke Learning Groups

The classification of motor skill learning in swimming teaching would facilitatean excellent learning outcome in the swimming course for children beginners, it willhelp identifying the different individual`s learning progress and considering theirunique aptitude accordingly. The scientific presumption of whether the physiquevariables could be used in forecasting the children`s swimming learning results is notonly meant to provide an effective reference for skills learning pre-classification inteaching, but also can be used to draw a preliminary understanding for the teachingobjects before the breaststroke teaching classes are initiated. Modern life bringsseriously negative effects on children`s Sensory Integration. By analyzing thecharacteristics of the SID （Sensory Integrative Dysfunction） rate measured fromgroups of different breaststroke learning effects as well as the changing characteristicsof children`s breaststroke learning outcomes in groups of different SensoryIntegration levels, this research help better understanding that whether physiologicaleffects of SID such as Motion Dysfunction, Spatial Perception Dysfunction, andVestibular Dysfunction could negatively influence the breaststroke skills learning. Byusing Methods of Interviewing, Testing, and Mathematical Statistics, this papermainly takes the prediction model of breaststroke learning outcome, its evaluationmethod and the influence of Sensory Integration level on breaststroke learning as theresearch objects, to establish the prediction model based on the experimental datumcollected from108trainees aged from6to11; besides, we analyze the feature ofchildren`s Sensory Integration level between groups of different breaststroke learningoutcomes as well as the features of children`s breaststroke learning outcomesbetween groups of different Sensory Integration level. The research results are asfollowed.1. A highly significant difference is found between male groups of differentbreaststroke learning outcome in terms of the Height of Raised Hand, Stature, BodyWeight, Sitting Height Index, Palm Area, BMI, Vital Capacity, Breath Holding Time,Standing Long Jump, Ratio of Shoulder Breadth to Stature, Height of VerticalJumping, Sit-up, Leg length; and a significant difference is found between malegroups of different breaststroke learning outcome in terms of the High Leg-lifting Running, Habitus（Somatotype） Index, Ratio of Arm Span to Stature. A significantdifference is found between female groups of different breaststroke learning outcomein terms of Somatotype Index, Arm Span; and the difference of the mean values ofStature, Single Arm Length, Shoulder Width are highly significant between femalegroups of different breaststroke learning outcome. These findings indicate that thesevariables showing different degrees of significant difference between groups ofdifferent breaststroke learning level, are available in establishing the predictionmodels via appropriate ways.2. By One-way Analysis of Variance, Independent Sample T-test, andTwo-sample Non-parametric Test, this study finds that breaststroke learners who areolder and taller would have got a relatively excellent breaststroke learning outcomes,compared with these who are not. The mean values of Stature, Age in the excellentlearning group, common group and bad group are declining accordingly. This findinghad been further proved in the Analysis of Covariance, when precludes the influenceof the covariance, namely the Age and the Stature, variables which has highcorrelation with the covariances are to some extent unable to reflect the differentlevels of breaststroke skills learning under the circumstance of any randomized twocontrol variables of the three, thus Age, Stature are considered to be one of the mainfactors that influence the outcomes of breaststroke learning.3. The Principal Component Model is established by variables that hassignificant difference of different degree between groups of different breaststrokelearning outcomes, these variables are tested by One-way Analysis of Variance,Two-samples Independent T-test, and Two-sample Parametric T-test as to know theirsignificant levels before model establishment. And further prediction models are builtbased on the Principal Component Model, namely the Predicting Factor Model andthe Factors Regression Model. The Predicting Factor Model is aimed to forecast thechildren`s breaststroke learning outcomes based on their pre-tested variablesreflecting the physical conditions. The evaluation and prediction are processed bycalculating the aggregate score of factors. The mean value of the aggregate factorscore of the physique variables in boys`excellent group, common group, bad groupare5.95±4.77,-1.39±3.65,-4.21±3.93respectively, while these of girls`are1.71±3.30,-0.42±2.20,-1.16±3.59. The method of predicting a trainee`s expectedbreaststroke learning outcome is to calculate his or her aggregate factor score by thefactor score functions, and the breaststroke learning group that has the nearestdistance from the mean value of the group`s aggregate factor score to the trainee`s aggregate score is the prioritized consideration of deciding the trainee`s predictedgroup （the Nearest Neighbor）. The aggregate score is acquired by using thesubordinate factors score equations, and above all an aggregate factor score equation.The accuracy of Factor Prediction Model in boy`s excellent, common, and bad groupare80.00%、48.00%、70.00%respectively, a64.62%in average accuracy. Theprediction in girls`excellent, common, bad group are76.92%,23.53%,53.85%inaccuracy respectively, and the average predicting accuracy is48.84%.The Factor Regression Model is built based on the principal components（independent variables） extracted from the original physique variables of both boysand girls, taking the Z score reflecting the breaststroke learning outcomes as thedependent variable. The measure of prediction is to acquire a trainee`s predicted Zscore by regression equation, and then check the mean value chart, the group whichhas the nearest value distance between its mean value and the trainee`s predicted Zscore is highly considered in breaststroke learning outcome prediction. By retest usingthe original datum of108children, it can be found that the prediction accuracy forexcellent learning group, common group and bad group are55.00%,84.00%,60.00%,and a67.69%in average, as for girls, it is23.08%,88.24%,38.46%respectively, anda53.49%in average.4. By using the SPSS17.0statistical software, stepwise discrimination functionare built based on children`s physique variables （discriminating variables）. For malesdiscrimination function based on the body shape variables: F=0.44×ShoulderWidth-15.252, and the one based on the body functions variables: F=0.002×VitalCapacity-2.884, the functions based on the physical quality variables: F₁=+0.12×Sit-up+0.06×Squat and Up-5.081, F₂=-0.09×Sit-up+0.10×Squat andUp-1.329;for females discrimination functions based on the body shape variables: F₁=0.069×Arm Span+0.053×Habitus（Somatotype） Index-11.959, F₂=0.080×ArmSpan-0.037×Habitus（Somatotype） Index-9.128. When making a prediction is wanted,take the variables of what has been described in the functions into the model neededand then get the values. Group with the nearest mathematical distance from the groupGravity Center （the mean value）to the value calculated is the predicting result inwhich the trainee is belonging to. The accuracy of predicting by discriminationfunction based on the body shape variables in boys`excellent, common, and badgroup are80.00%、44.00%、50.00%respectively, and a56.92%in average, theaccuracy of the one based on the body functions variables in boys`excellent, common,and bad group are60.00%,32.00%,75.00%, and53.85%in average, and are65.00%, 28.00%,65.00%if predicted by physical quality variables, with a then50.77%accuracy in average; while girls`prediction in accuracy by body shape variables are53.80%、41.20%、61.50%respectively, with a51.16%in average.5.Compared with the accuracy of each three predicting model described above,the highest average accuracy of predicting for excellent group in male trainees is80.00%, which are predicted both by Factor Prediction Model and Factor RegressionModel, while is76.92%in female which is forecasted by Factor Prediction Model.Factor Regression Model has the highest efficiency in predicting the male`s commongroup, with a84.00%in accuracy, while the same is in female as Factor RegressionMode has a accuracy of88.24in predicting the common group. The discriminationfunction based on male`s body functions has the best prediction results in accuracyfor male`s bad group, with the accuracy as high as75.00%, while the discriminationfunction based on female`s body shape variables take the first position in female`sbad group with a accuracy of61.50%. The Factor Regression Model are both superiorin male`s and female`s general prediction accuracy, with the accuracy of67.69%and53.49%respectively.When a prediction is initiated, lay out the results by these models describedabove, model with highest accuracy should be prioritized as the predicting results maybe different. For example, if a boy`s breaststroke learning outcome is wanted beforeteaching, get the prediction results by various models according to their rules. It ismay be shown the boy should belong to excellent group by Factor Prediction Model,and should be in bad group if predicted by Factor Regression Model, but the resultfrom Factor Regression Model should be highly concerned as to its higher predictingaccuracy（60.00%） for bad group.6. There is no significant difference between males and females in SID rate, nosignificant difference in breaststroke skills learning outcomes between group withnormal sensory integration and groups with different degrees of SID, and in SID ratebetween groups of different breaststroke learning results, thus it can be deduced thatchildren`s breaststroke skills learning outcomes has nothing to do with their SensoryIntegration level. By One-way Variance Analysis, it can be drawn that the controlvariables, namely the different levels of Sensory Integration, has no significantinfluence on children`s breaststroke learning outcomes. These findings suggests thata relatively excellent breaststroke skills learning could also be achieved even thoughthe child`s sensory integration level is slightly or seriously lower than the normals.Conclusion. 1. There are significant differences between groups of different breaststrokelearning outcomes in some of the physique variables in children trainees, thus thesevariables could be selected to establish the prediction models of breaststroke learningoutcomes.2. One of the main factors influencing child`s breaststroke learning outcomes isAge and Stature.3. The highest average accuracy of predicting for excellent group in maletrainees is80.00%, which are predicted both by Factor Prediction Model and FactorRegression Model, while is76.92%in female which is forecasted by FactorPrediction Model. Factor Regression Model has the highest accuracy in predicting themale`s common group, with a84.00%in accuracy, while the same is in female asFactor Regression Mode has a accuracy of88.24. The discrimination function basedon male`s body functions has the best prediction results in accuracy for male`s badgroup, with the accuracy as high as75.00%, while the discrimination function basedon female`s body shape variables take the first position in bad group with a accuracyof61.50%. The Factor Regression Model are both superior in male`s and female`sgeneral prediction accuracy, with the accuracy of67.69%and53.49%respectively4. Child suffering SID could also achieve an excellent breaststroke learningoutcome.