• Bhagwati Galande
• Prof. M. K. Kodmelwar

Nearest neighbour (kNN) methodology could be a widespread classification methodology in data processing, classification and statistics due to its easy implementation and significant classification performance. However, it's impractical for ancient kNN strategies to assign a k to any or samples. Previous solutions of classification assign completely different k values to different samples by the cross validation methodology however area unit sometimes time intense or time consuming. This paper proposes a kTree methodology to be told completely different best k values for various test/new samples, by involving a coaching stage within the kNN classification and Particle Swarm improvement (PSO). Specifically, within the coaching stage, PSO improvement is employed to search out best k values and discarded week samples then kTree methodology 1st learns best k values for all coaching samples by a brand new distributed reconstruction model, then constructs a tree namely, kTree. kTree quick outputs the best k worth for every sample, and then, the kNN classification is conducted. As a result, the planned kTree methodology features a similar running value however higher classification accuracy, compared with ancient kNN strategies, that assign a fixed k to any or all samples.

"Efficient Classification for Dataset Using PSO and KNN Implementation", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 6, page no.213-216, June 2019, Available :http://www.jetir.org/papers/JETIR1906L33.pdf

"Efficient Classification for Dataset Using PSO and KNN Implementation", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 6, page no. pp213-216, June 2019, Available at : http://www.jetir.org/papers/JETIR1906L33.pdf