Machine learning Methods for Performance Forecast and Assessment of Female Handball Players

Abstract

System learning algorithms are used to carry out tasks that people find difficult to do. The evaluation and forecasting of the performance of precise fitness sports using players as a resource is becoming more and more important in both league and practice scheduling. When it becomes Due to the use of conventional methods, the variety and difficulty of particular sorts of sporting events, and the usually time-variant interactions among them, sports are difficult to research and predict. Strong Machine Learning (ML) algorithms can analyse gamers' physical requirements with remarkable accuracy

In order to develop a more effective tool and uncover the key variables affecting projected results in female handball athletes, this study aims to compare several machine learning (ML) approaches to predict specific participant achievement kinds. The simple type of regression in machine learning (ML), i.e. Simple Linear Regression (SLR), Classification Tree (CT), Support Vector Regression (SVR), and Neural Networks that employ Radial Basis Function (RBFN), were used to predict the performance abilities of female handball players in the Squat Jump (SJ), Squat Jump on Toes (SJT), Sprint over a ten-m distance (SP10), and a Handball Sport-Skill Test (HSST). For each ML version, 117 occurrences of training samples with a maximum of 23 feature values have been recorded.

With R-squared values ranging from 0.86 to 0.97, the results showed that the RBFNN outperformed superior models and switched from red to green when forecasting players' performance. Using numerous common performance criteria, such as mean squared errors (SE) and implied absolute errors, we also assessed all of the fashions (AE). Finally, by upgrading the superior instrument, important factors affecting expected achievement had been assessed. The results are encouraging and intriguing for future researchers, despite this being the earliest and most preliminary attempt to use ML in the problem of sports, namely handball.