Dimensionality Reduction of Fuzzy Soft Sets using FPSO algorithm

Abstract

Nowadays, soft set theory based techniques have been drawing some attention for decision making in various fields. Especially the most significant one is soft sets based parameter reduction, an important topic for both artificial intelligence and information science research community. The major intention of parameter reduction technique is to eliminate the set of parameters that describe a specified soft set without varying its principal choice objects. The paper utilizes the representation of fuzzy soft sets by fuzzifying soft sets. The work analyses certain distinct fuzzy soft sets based parameter reduction algorithms. The primary goal is to compare Fuzzy-PSO and normal PSO algorithm to highlight their individual performance with respect to computational complexity and parameter reduction efficiency. These algorithms are compared to traditional algorithms also to verify their computational complexity and parameter reduction ability. Finally the paper proposes FPSO technique as preferred algorithm for general datasets. The survey is useful especially in dimensionality reduction of datasets with large number of parameters. The paper shows discussions on KDD dataset used in the research of Intrusion detection techniques.