• Dr. D. Mabuni

The impact of artificial intelligence and machine learning is very high on the society and in particular Machine Learning applications usages are increasing rapidly everyday and they became indispensible with the daily activities of human beings. Accurate learning models are very much useful nowadays and in particular decision tree classification models are needed in diversified domain applications. The popular k-fold cross-validation technique is particularly useful in determining the predictive performance of the model. Previous research communities had already decided that k value in k-fold cross validation must always be 10 or 20 but in real time situations it is not always true. That is sometimes k value may be 5 or 6 or 15 and so on. In the present study by employing many standard and UCI machine learning datasets experiments are conducted. After inspecting and observing experimental results thoroughly new postulates such as k value in k-fold cross validation may be taken into consideration other than usual k = 10 or k = 20 in some applications. That is, always k value should not be 10 or 20 instead k may take on any positive integer (k > 1) value.

optimal k value, k-fold cross validation, k = 10 or k = 20, predictive performance, decision tree classification, machine learning models

"A New Framework for Finding Optimal k Value in k-fold Cross Validation Technique in Machine Learning Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 6, page no.k47-k54, June-2023, Available :http://www.jetir.org/papers/JETIR2306A05.pdf

"A New Framework for Finding Optimal k Value in k-fold Cross Validation Technique in Machine Learning Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 6, page no. ppk47-k54, June-2023, Available at : http://www.jetir.org/papers/JETIR2306A05.pdf