This paper makes clear about two popular algorithms namely k-Nearest Neighbors (K-NN) and Condensed nearest neighbor (CNN) and accomplishes a comparison between them; nowadays classification is a significant in data mining methodology. It can be used for classifying the interested users. The classification algorithms have been projected in the earlier eight decades. Every part of them has their own highlights constraint and highlights limitation. All of them have changeable run time, space necessities, consistency and ability for higher performance. We explore the effect of a level of consciousness on the "nearest neighbor" problem. It also talks about parameter selection, the I-nearest neighbour classifier, nearest neighbor search and various methods in it. This paper also concludes about CNN for data reduction.