The Research And Application Of Clustering Analysis And Recommendation In Sports Competition Stress

With the development of modern sports game, the quantity and size of sports increases so quickly. Athletes have a strong willing to be high level player on their field, pressure is an important factor that must be overcome. As a special type of pressure, competitive pressure mainly arises from the competition environment. It has a negative impact for our athletes. As a consequence of these, more and more experts pay attention to competition pressure.Because of the variable type of stress suffering from athletes, many relevant questionnaires and scales are developed by psychology experts for athletes’daily life. Moreover, psychology experts collect a lot of pressure source data and analysis these data.Thus our athletes can have a healthy body and mind to make a good achievement on the playground.In this article we try to research the pressure source data using natural language processing and machine learning. Focus on two commonly used clustering algorithms and recommendation algorithms. From this point of view, we can get some new research result. First, we introduce all of the sports pressure questionnaires that are used in this paper, processing the source data, normalizing them and so on. Second, we introduce the fundamental concepts of clustering algorithms. For the features of sports pressure source data, we addresses an important issue about time complexity and cluster quality and try to improve the traditional K-Means algorithm. We compare the old algorithm with the new improved hierarchical K-Means algorithm. Experiment results show that the new algorithm is efficient in reducing time complexity. Thus we apply the new algorithm to our sports pressure source data. Finally, in order to help athletes with a good advice and keep them a healthy body, we construct a model with athletes’ sports pressure source data, then content-based recommendation is used to identify the useful strategies.