(Retracted) Image analysis and pattern recognition method of three-dimensional process in physical education teaching based on big data

Abstract

There are some irrelevant or low correlation features in physical education teaching pattern recognition, resulting in low recognition accuracy. A three-dimensional integrated process physical education teaching pattern recognition method based on big data is designed. The physical education teaching data are cleaned, integrated, normalized and discretized, and the data attribute features are extracted using big data technology. The results of an attribute in all lifting decision trees are weighted and averaged to obtain the characteristic importance score of the attribute. The teaching pattern classification algorithm is designed according to the attribute characteristics, and the trace of the sample divergence matrix is used to increase the speed of solving the parameters and weights. Through the preliminary and secondary training of neural network, a physical education pattern recognition model is established to ensure the accuracy of small sample feature set recognition. It shows that the recognition accuracy of this method is significantly higher than that of the recognition methods based on the random forest algorithm and convolution neural network, and it has a good performance in physical education teaching pattern classification and recognition.