Investigation of Random Forest Performance with Cancer Microarray Data

The diagnosis of cancer type based on microarray data offers hope that cancer classification can be highly accurate for clinicians to choose the most appropriate forms of treatment with it. Due to several inherent characteristics associated with microarray data, accurate diagnosis has been an active research topic attracting tremendous research interests in machine learning community. In this paper, random forest classifier is applied to a cancer microarray data in an attempt to achieve more accurate and reliable classification performance. Impact of gene reduction to classification rates was evaluated and an attempt was made to identify a gene selection method which uses small number of genes, yet yield a high classification rate. Random forest performance in microarray data classification in general was also investigated.