• Dr. D. Mabuni

In data analytics K-nearest neighbor classification algorithm is given much importance for data classification. Finding optimal K-value and then for a given test tuple searching for all the K-nearest neighbors are two separate and critical tasks in K-nearest neighbor classification algorithm. Many methods exist for optimal K-value determination including cross validation technique. In this paper a new technique is proposed for optimal K-value determination using decision tree classifier model constructed from the given training dataset. Optimal K-value is determined directly by counting the number leaf nodes in the decision tree classifier. That is K-value is set equal to the number leaf nodes of the decision tree. The most important advantage of the decision tree classifier is that without determining K-value separately it is possible to find class label of the test tuple easily by searching and finding the correct leaf node in the decision tree and then setting the class label of the leaf node as the class label of the test tuple. In case if K-value is required then use proposed heuristic procedure, number of leaf nodes of the decision tree is set as K-value. Also, experimentally verified that the optimal K-values determined through K-nearest neighbor classification algorithm are almost equal to the K-value determined through the proposed decision tree classifier technique. Decision tree classifiers are powerful, well established, bench marking standard, highly accurate, scalable, and predominantly using techniques in both machine learning and data mining.

Optimal K-value, K-value, decision tree classifier, multiple search trees, K-nearest neighbour values, KNN, K-nearest neighbor classification, machine learning, data mining.

"A Heuristic Optimal K-value Determination in K-Nearest Neighbor (KNN) Classification using Decision Tree Technique", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 9, page no.f542-f550, September-2023, Available :http://www.jetir.org/papers/JETIR2309566.pdf

"A Heuristic Optimal K-value Determination in K-Nearest Neighbor (KNN) Classification using Decision Tree Technique", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 9, page no. ppf542-f550, September-2023, Available at : http://www.jetir.org/papers/JETIR2309566.pdf