Team strategizing using a machine learning approach

Abstract

Team strategizing is an important aspect which requires critical analysis to ensure a desirable near-optimum performance. The key to solve this issue is by tapping the available talent within the team which at times, can be elusive. With increasing competition, a talented team, with an ineffective and outdated scouting strategy, may have to face unfavourable results. In this paper, we have conducted research in the domain of Sports, specifically Soccer. Strategy considered in the research is centered around deciding the lineup of a team by assessing the skillset of the players. Considering the novelty of the approach, we have developed our own web scraping algorithm to collect the dataset. Machine Learning models like Neural Network(MultiLayer Per-ceptron), Random Forests and Logistic Regression have been used to make predict the position a particular player will perform best at. The accuracy of the said models have been analysed for comparative analysis.