A hybrid Artificial Neural Network-Genetic Algorithm approach for classification of microarray data

Classification of micro-array data which are used in diagnosis of cancer studies is one of the important topics in bioinformatics field. In these studies, an excessive number of features of micro array data increase the data dimensions. This situation makes it difficult to analyze data with conventional algorithms and approaches. In this study, dimension reduction is applied to micro array data generated from tissue obtained from patients with have CMS (Central Nervous System) tumors. After that, this data is classified with Artificial Neural Networks strengthened with Genetic Algorithms. The findings are compared to finding obtained from simple and powerful classification algorithms such as KNN (K-Nearest Neighbors) and SVM (Support Vector Machine). Highest performance value is found 80% certainty, sensitivity 71.4%, 75% accuracy rate through Artificial Neural Networks strengthened with Genetic Algorithms. Obtained results show that strengthening of Artificial Neural Network with Genetic Algorithms provides higher performance classification.