Comparison of Methods for Automated Feature Selection Using a Self-organising Map

摘要: The effective modelling of high-dimensional data with hundreds to thousands features remains a challenging task in the field machine learning. One key challenges is implementation methods for selecting set relevant features, which are buried along irrelevant noisy by choosing subset complete input that predicts output higher accuracy comparable performance set. Kohonen’s Self Organising Neural Network MAP has been utilized various ways this task. In work, review appropriate application multiple carried out. feature selection approach based on analysis network result after training presented comparison two methods.