Type-specific classification of bronchogenic carcinomas using bi-layer mutated particle swarm optimisation

The cancer disease is posing a big challenge in the field of pathological diagnosis. The feature selection of cells is highly important in isolating the affected cells. The classification of cancer cells is gaining importance among clinical researchers. Gene expression profile (GEP) is used in better classifying genes in a cell or tissue. Gene expression data (GED) differs for every gene from which cell or tissue it is originated. Based on the GED the cancer cells can be classified into seven categories from which cell or tissue it was born. The infected cells can be graded from level one to four based on its growth and difference from other unaffected cells. Many techniques have been developed in the past for classifying cancer affected genes. In this paper we propose a modified classification algorithm bi-layer mutated particle swarm optimisation (BLMPSO). The microarray dataset used for testing the method is Affymetrix Human Genome U95Av2 Array. The simulation results showed that the proposed technique performs better in terms of classification based on GED than the other existing methods.