Performance investigations of various clustering algorithms for microarray gene data

Abstract

Microarray gene data is a potential tool for analysing the gene expressions of living organisms. However, the microarray technology is highly challenging to process, owing to its volume. The microarray technology based decision making systems are quite promising than the systems based on medical images. All the worldly living organisms are based on the basic functional units called genes and when the status of genes is tracked, it is simple for the healthcare experts to make clinical decisions. newlineUsually, the microarray gene data analysis is performed in diagnosing and grading cancer. The behaviour of genes is studied and the decisions are made accordingly. This kind of analysis is more reliable than the traditional approaches, however the systems based on microarray data are quite scarce in literature. newlineThe microarray gene data can be analysed in two ways, which can either by supervised or unsupervised approaches. The supervised approach is also called as classification, which requires adequate training for differentiating between the data samples. The process of training involves a set of data samples along with the classes to which the samples belong to. newlineThe unsupervised based analytical approaches involve no training process and this makes sense that the analysis can be done on-the-go. The goal of this research work is to analyse the performances of different versions of Fuzzy C Means (FCM) with different meta heuristic algorithms. newline