With the advances of computers and computing methods, researchers have started to use more advanced analysis methods, such as principal component analysis, support vector machines, and regressions on larger datasets 1– 8. Most biomechanics studies with large numbers of data and researchers use educated guesses for data reduction before starting data acquisition, and they are limited to investigation of only several variables at specific events (e.g., shoulder flexion angle at ball release). Therefore, due to the nature of the biomechanics research, most studies have had a highly limited number of trials, which makes it hard to use advanced data analysis methods, such as machine and deep learning.
Three-dimensional motion analysis systems, which use markers to track motion, are expensive, and they require long amounts of time.
