Ruijie Li and  ChaoZheng. Sports Performance Test Based on Regression Coefficient Under Ordinary Least Square Method. Dynamic Systems and Applications 29 (2020) No. 5, 2100 – 2109

https://doi.org/10.46719/dsa202029544

ABSTRACT.
Professional sports organizations, to invest considerable resources to understand the factors that influence the better performance, the data is collected, has been analyzed. All the latest Advances in non-invasive technologies such as global positioning system (GPS) do the average coaching does not have, and sports scientists readily available with a large amount of data. However, analysis of such data, mainly the sample size is small, it may become difficult if there are more than one highly correlated variables into a dataset. This article aims to overcome the problems multicollinearity shows how you can use a load of training (TL), partial least squares method, about our analysis of the proposal (PLSCA) method, to identify a new “one variable out of the false” (LOVO). This variable affects the most “exclusive” players at a young rugby league. If you quantified in perception overwork (SRPE), – men 16 points of professional rugby league players, and player session evaluation at six weeks of pre-season training limited eyes were accumulated TL… Immediately before and after the training period Participants received a 30-15 interval fitness test (30-15IFT).To decide the start of fitness, “exit Fitness.” In the collected TL variable of a total of 12, we have a “smart fitness” as a covariate, which is the opposite direction to the “high-end fitness.” However, multiple linear regression (MLR) process means that it is unstable data (VIF> 1000 9 single variables), and another equivalent multicollinearity quantifies the relative importance of predictors has developed a new LOVO PLSCA adapted for.It has been determined to be the most critical TL variables that affect the improvement of the physical fitness of the players. This variable brings the singular value of the inertia (5.93); it will be the most significant decrease. Suppose it is included in the complete linear regression model. In that case, the variables are used as a covariate “Start compatible,” v30-15IFT seventy-three percent of the variance of the “end compatible” (P <0.001) describes a plurality of to eliminate the problem.If you want to use it as a filtering tool to generate the MLR model LOVO PLSCA “it is a very cumulative distance at high speed” to prove if it contains a “smart fitness” as a covariate in the MLR model, predictors It’s essential. In this way, LOVO PLSCA is for sports scientists and coaches to try to analyze the data set using GPS and MEMS technology, which can be a useful tool.