Selecting significant marker genes from microarray data by filter approach for cancer diagnosis

• DOI: 10.1016/j.procs.2018.01.126
• Publication Date: 2018-03-18T18:05:48Z
• Authors: Sara Haddou Bouazza; Khalid Auhmani; Abdelouhab Zeroual; Nezha Hamdi
• Keywords: 0303 health sciences; 03 medical and health sciences; 3. Good health

In machine learning, feature subset selection phase is the process of selecting a small most relevant features for use in model construction. The main goal this paper to perform comparative study between methods applied DNA microarray dataset and investigate strength each method. studied are: F test, T Signal noise ratio (S/R), ReliefF Pearson product-moment correlation coefficient (CC). This carried out using five cancers; Leukemia, Lung, Lymphoma, Central Nervous System Ovarian cancers. Evaluation has been done by supervised classifiers: K Nearest Neighbors (KNN), Support Vector Machines (SVMs), Linear Discriminant Analysis (LDA), Decision Tree Classification (DTC) Naïve Bayes classifier (NV). purpose classification predict presence cancer. accuracy measured selected genes. Results show that combination S/R method KNN present highest accuracies different 100% only 13 genes Leukemia cancer, 21 Lung 4 Lymphoma 76.7% 6 CNS 30 ovarian