Research on basketball sports neural network model based on nonlinear classification

Intelligent video analysis has broad application prospects. How to automatically analyze and identify human behavior in video has attracted extensive attention from researchers at home and abroad. Moreover, researching effective video behavior recognition algorithms and designing efficient behavior recognition systems has important theoretical and practical value. This paper studies the nonlinear classification technique and applies the video behavior recognition algorithm to basketball recognition. Moreover, this paper studies the classical convolutional neural network model and several improvements. In addition, this paper explains the advantages of convolutional neural networks in feature extraction compared with traditional neural networks and analyzes the performance of the algorithm by designing actual experiments. The research results show that the algorithm can quickly identify multiple players on the field, and the method can effectively deal with occlusion and other issues with high accuracy and real-time.

Basketball sports neural network model based on nonlinear classification

With the continuous innovation of science and technology, the mathematical modeling and analysis of bodily injury in the process of exercise have always been a hot and difficult point in the research field of scholars. Although there are many research results on the nonlinear classification of the basketball sports neural network model, usually only one model is used, which has certain defects. The combination forecasting model based on the ARIMA model and neural network based on LSTM can make up for this defect. In the process of the experiment, the most important is the construction of the combination model and the acquisition of volunteer data in the process of the ball game. In this experiment, the ARIMA model is used as the linear part of the data, and LSTM neural network model is used to get the sequence of body injury. The results of the empirical study show that: it is reasonable to divide the injury of thigh and calf in the process of basketball sports, which is very consistent with the force point of the human body in the process of sports. The results of the two models predicting the average degree of bodily injury for many times are about 0.32 and 0.38 respectively, which are far less than 1. The execution time of the program for simultaneous prediction on the computer is about 1 minute, which is extremely effective.

Measuring Sequences of Representations with Temporally Delayed Linear Modelling

SUMMARYThere are rich structures in off-task neural activity. For example, task related neural codes are thought to be reactivated in a systematic way during rest. This reactivation is hypothesised to reflect a fundamental computation that supports a variety of cognitive functions. Here, we introduce an analysis toolkit (TDLM) for analysing this activity. TDLM combines nonlinear classification and linear temporal modelling to testing for statistical regularities in sequences of neural representations. It is developed using non-invasive neuroimaging data and is designed to take care of confounds and maximize sequence detection ability. The method can be extended to rodent electrophysiological recordings. We outline how TDLM can successfully reveal human replay during rest, based upon non-invasive magnetoencephalography (MEG) measurements, with strong parallels to rodent hippocampal replay. TDLM can therefore advance our understanding of sequential computation and promote a richer convergence between animal and human neuroscience research.