BPSO Optimized K means Clustering Approach for Medical Data Analysis

Data mining plays a very important role in the analysis of diseases and clustering approach makes it easier to classify the data collected in respective groups. Medicine companies and medical appliance manufacturer are benefitted from these data analysis. Now a days, this is done at a very large scale and has been named as big data analysis in which data size is of many terabytes. Optimization forms an integral part of our day to day life. It can be defined as an art of selecting best alternative from a set of options. Several global optimization algorithms have been developed. PSO (Particle Swarm Optimization) is a powerful optimization technique. It consists of a population of solutions called as particles where the position of particles is determined on the basis of position vector and velocity vector. The position of particles gets changed in search of optimal solution. The particles are distinguished as personal best and global best. Hybrid algorithms combine the desirable properties of different algorithms to mitigate weaknesses of individual algorithms and result in optimal solution. For example PSO combined with GA, DE and results in DE-PSO and GA-PSO which are better versions of PSO. Bacteria foraging optimization algorithm is another type of optimization algorithm which is based on the behavior of biologically inspired E-Coli bacteria. EColi bacteria search the search space for rich nutrients by using their energy per unit time. The common characteristic bacteria are grouped together. The bacterium communicates with each other by sending signals. The BFO is used by many researchers recently and they try to hybridize the BFO with different algorithms to find the local best and global best solution in the search space. In this paper, BPSO, a hybrid algorithm made up of BFO and PSO algorithms uses k means clustering approach for making clusters to obtain an optimal solution.