Breast cancer risk assessment prediction using an ensemble classifier

Breast cancer is still the most common killing disease nowadays among women. The current research is to identify accurately the risk of breast cancer. Many machine learning models can be used to predict the risk based on gene expression data. When analyzing the gene expression data, the key concern observed is how to select the best informative set of genes. Usually, the number of gene attributes far exceeds the number of data observations. This problem warrants a feature selection algorithm to select a subset of suitable genes. In this paper, an ensemble classifier with Correlation based feature selection with forward search is proposed for use in microarray breast cancer gene expression datasets. The proposed ensemble classifier helps to classify the relapse with the most informative genes, which can help physicians to identify breast cancer at the early stages. The evaluation reveals that the proposed method achieves higher classification accuracy, compared to existing works.