Comparative Study on Feature Selection Methods to Reduce High Dimensionality in Big Data

Abstract

Big information may be a combination of structured, semi structured and unstructured information collected by organizations that may be strip-mined for info and employed in machine learning comes, prognosticative modeling and different advanced analytics application. Spatiality in statistics refers to what number attributes a dataset has, care information is ill-famed for having Brobdingnagian amounts of variables in a perfect world; this information may be depicted in a very unfold sheet, with one column representing every dimension. In observe, this can be tough to try to, in past as a result of several variables square measure inter-related (like weight and blood pressure).This paper gift study on feature choice technique to cut back high spatiality issue in huge information.