Optimization-Based Effective Feature Set Selection in Big Data

Abstract

Of late, the data mining has appeared on the arena as an ideal form of knowledge discovery crucial for the purpose of providing appropriate solutions to an assortment of issues in a specified sphere. In this regard, the classification represents an effective method deployed with a view to locating several categories of anonymous data. Further, the feature selection has significantly showcased its supreme efficiency in a host of applications by effectively ushering in easier and more all-inclusive remodel, augmenting the learning performance, and organizing fresh and comprehensible data. However, of late, certain severe stumbling blocks have cropped up in the arena of feature selection, in the form of certain distinctive traits of significant of big data, like the data velocity and data variety. In the document, a sincere effort is made to successfully address the prospective problems encountered by the feature selection in respect of big data analytics. Various tests conducted have upheld the fact that the oppositional grasshopper techniques are endowed with the acumen of effectively extracting the requisite features so as to achieve the preferred outcome Further, enthusing experimental outcomes have revealed the fact only a trivial number of hidden neurons are necessary for the purpose of the feature selection to effectively appraise the quality of an individual, which represents a chosen subset of features.