This experiment explores the Genetic Algorithm for our scouting model. This method has the potential to assist clubs with a smaller budget find players capable of performing at a high level to increase parity throughout leagues around the world. Genetic algorithm is a heuristic optimization technique that induces variability amongst populations and creates a pool of possible outcomes to find the best solution
Below image shows the process of mutation and crossover in our dataset
Each player is assigned a fitness score by the fitness function, which evaluates the ability of the player to compete with other players. Player selected for the next iteration is based on its fitness score. Fitness function has been modified to add the single most important attribute for each position to narrow down to specific players with relevant skillset for that position. The most influential characteristic of a player for a specific position is chosen as the attribute for identifying the best players.
Experiment Result: