A Hybrid Feature Selection Technique based on TOPSIS and Rough Sets for Cancer Classification in Intuitionistic Fuzzy Normalized Microarray Data

In microarray gene expression data, the performance of classification tasks highly depends on the discriminative, relevant and informative features because of the high dimensionality of microarray data. Feature selection algorithms are used for extracting a more precise and compact set of features which improves the classification accuracy and efficiency.
This paper proposes the hybrid feature selection technique based on Techniques for Order Preference by Similarity to an Ideal Solution (TOPSIS) and Rough sets to handle imprecise and vagueness of features in the data set. This work starts with the Intuitionistic Fuzzification as a primary step which has been proven to be an effective and efficient tool in the process of selecting relevant and non-superfluous features. The Intuitionistic fuzzified data is fed into the TOPSIS method to select the set of discriminative features. The selected features are given as input to the Rough set technique to select the final set of significant features. The Rough set based rule generation and prediction is performed with the selected set of features.