Microarray gene selection using self-organizing map

Accuracy, precision, and rapidity of disease prediction are important for disease evaluation in clinic and laboratory studies because different diseases would have different drugs and treatments. This study presents a new technique for cancer prediction from DNA microarray data. The prediction composes of two main steps that are the step of the important gene selection by using statistic methodology and the step of clustering cancer data by using self-organizing map. The experimental DNA microarray data sets are carcinoma, leukemia, and lung cancer. The experimental results are the rules of gene with 100% accuracy for cancer prediction.