Identifying High-Risk Breast Cancer Patients Using Microarray and Clinical Data

The performance of DNA microarray in breast cancer prediction demonstrates the potential of genome-wide analysis using gene expression data. Therefore, this study proposed a prediction method called GridPCA to identify highrisk breast cancer patients using both microarray and clinical data. The GridSearch and Principal Component Analysis are employed in the proposed method to deal with the high dimensionality of microarray data. The experimental results showed that GridPCA achieved approximately 82% of average predictive accuracy with Decision Tree, K-Nearest Neighbour, Logistic Regression and Support Vector Machine classifiers. In future, the proposed method could be used in developing systems that help doctors in planning, decision making and tailoring appropriate treatments for increasing the survival rate of breast cancer patients.