Data Processing for Large Database using Mapreduce Approach and Using APSO

Big Data is a hype up-springing many technical challenges that confront both academic research communities and commercial IT deployment. It is generally known that data which are sourced from data streams accumulate making traditional batch-based model induction algorithms. Feature selection has been popularly used to lighten the processing. Optimal feature subset which is derived grows exponentially in size. In order to tackle this problem, a novel lightweight feature selection is proposed and it is designed particularly for mining streaming data on the fly, by using accelerated particle swarm optimization (APSO) which is achieved with good accuracy also with reasonable processing time. In this paper, the data in the disk are processed in subsequent iterations.