IDENTIFYING CIRCULATING TUMOR CELLS IN BREAST CANCER WITH DATA MINING ALGORITHMS BY USING MICROARRAY

Breast Cancer (BC) is an extremely diverse disease and extremely widespread among western women. Circulating tumor cells (CTCs) in Peripheral Blood (PB) is one of the most significant diagnostic factors for the cause of BC, the genomic study regarding CTCs detection in PB specifically for BC is limited because of lack of genes features for their identification and separation.  As an alternative of direct CTC detection methods, in this study, we majorly focus on the Heterogeneous Swarm intelligent based Clustering Ensemble Framework (HSCEF) for the detection of distant factors in Peripheral Blood (PB). Proposed HSCEF combines the procedure of three clustering methods such as Hierarchical Levy Flights based Firefly Algorithm (HLFFA), Hierarchical Modified Artificial Bee Clustering (HMABC) and Semi-Supervised Clustering (SSC) which classify the selected gene features into Meta Static (MS), Non Meta Static (NMS), MS and NMS. In the proposed HSCEF, the similarity measurement results of various optimization methods are fused into single metric depending on Weighted Quality (WQ), which in turn to improve CTCs detection results in Peripheral Blood (PB). Publicly accessible breast cancer and PB microarray datasets is used in this work for experimentation of HSCEF procedure to the detection of several discriminant factors in Peripheral Blood (PB). Experimentation results is conducted to GSE29431 dataset samples and evaluated using the classification parameters like precision, specificity and classification accuracy.