Cheerleading athlete's action safety in sports competition based on Kohonen neural network

Skill cheerleading is a sport with high difficulty, high skill and relatively high injury probability. This study mainly discusses the judgment of Cheerleading athletes' action safety in sports competition based on Kohonen neural network. The Kohonen network consists of an input unit layer and a two-dimensional output network of processing units. During the training process, each element competes with other units to obtain each record. When a yuan obtains a record, its weight is adjusted to more closely match the predicted category of the record. The Kohonen neural network's analysis of data is divided into roughly two processes. One is the tentative evaluation process of the network model to obtain the overall pattern of the data, and the other is the process of adjusting and optimizing the network to obtain a better model for the features contained in the data. Athletes' injuries are not only caused by their own factors, but also caused by technical factors, poor protection and improper cooperation. Of course, skill cheerleading is not only difficult, but also includes many people participating in the completion of transition connection and some single person dance combinations, hand position combinations, jumps, etc., which may also become potential factors for athletes' injury. According to the characteristics of skill cheerleading team, the injury of athletes is divided into three parts. The first part is the injury of multi person cooperative action unique to skill cheerleading team, including throwing, lifting, pyramid, top and base in transition and connection. The second part is the injury of difficult somersault completed by a single person. The third part is the factors other than difficulty, including jumping, dance combination and hand position combination. The fuzzy Kohonen clustering algorithm proposed in this study adopts batch processing for motion risk data samples, eliminates the dependence of clustering results on the order of input samples, and makes it suitable for dealing with the problem of fuzziness. Finally, the judgment matrix is introduced to judge the risk index weight of the action safety of skilled cheerleaders in sports competitions. The injury rate of skill cheerleading athletes is different, and the injury rate of somersault difficult movement is 67.92%. The injury rate of external factors such as jumping, dancing and hand position combination was 48.15%. This study will help to provide useful theoretical guidance for the standardized, scientific, sustainable and good development of skill cheerleading team.