A review of hybrid machine learning approaches in cognitive classification

Abstract

The classification of functional magnetic imaging resonance (fMRI) data involves many challenges due to the problem of high dimensionality, noise, and limited training samples. In particular, mental states classification, decoding brain activation, and finding the variable of interest by using fMRI data was one of the focused research topics among machine learning researchers during past fewdecades. In the context of classification, algorithms have biases, i.e., an algorithm perform better in one dataset may become worse in other dataset. To overcome the limitations of individual techniques, hybridization or fusion of these machine learning techniques proposed in recent years which have shown promising result and open up new direction of research. This paper reviews the hybrid machine learning techniques used in cognitive classification by giving proper attention to their performance and limitations. As a result, various domain specific techniques are identified in addition to many open research challenges.