The Big Data has too many challenges that face each IT deployment and educational analysis communities, the huge amount of data coming from various sources which are based on data stream and curse of dimensionality. The Big Data depends on 3V challenges specifically, Volume, variety and velocity. It's typically illustrious that the data coming from various data sources in different format and gather together in very high speed and creating ancient batch based model which is infeasible for real time data processing. This can be the most important challenge with the Big Data. As velocity is one of the challenges in Big Data, the crucial issue is to mine most valuable or actual and relevant information. To perform data mining over such high speed information the Big Data technology obtaining importance currently a days. The Feature selection technique is employed for data stream mining on the fly in big data. Feature selection has been widely used to minimize the process load in causing the mining information model. To achieve the query accuracy within minimum processing time and to reduce the processing load the accelerated particle warm optimization (APSO) is employed.