PFCM Approach for Enhancing Classification of Colon Cancer Tumors using DNA Microarray Data

Colon cancer stands as one of the most lethal diseases affecting the large intestine. While it primarily afflicts older adults, its occurrence is not constrained by age. Typically, it originates as benign cell growth within the colon, eventually progressing to cancer due to somatic alterations affecting gene expression. DNA microarray technology offers a standardized means of assessing the expression levels of numerous genes. This technology distinguishes tumors across various anatomical regions based on gene expression patterns. However, the sheer volume of microarray data presents challenges, particularly in dealing with the curse of dimensionality. In the first level, PFCM technique is employed to reduce the dimensionality of genes or features. To evaluate the effectiveness of these techniques, we applied four suitable classifiers as Non Linear Regression, Logistic Regression, Bayesian LDC and Detrend Fluctuation Analysis. The classifier known as Detrend Fluctuation Analysis, which incorporates the PFCM Feature Extraction technique, achieves the highest level of accuracy, specifically 88.71%.