Comparison Classifier of Condensed KNN and K-Nearest Neighborhood Error Rate Method
G. I. Onwuka, Wasiu Afolabi Babayemi, I. E. Etuk · 2012
Error rate of classification from condensed KNN, cross validation, K-nearest neighborhood error rate are applied to data of 228 pregnant women from 114 training set and the other 114 were the testing set used to estimate the true error rates of the two classification method. In the two procedure, k-nearest neighborhood k=3 and above has the lowest error rate of 0.439. Since, the lower the error rate, the better the procedure. Therefore the k-nearest neighborhood (KNN) gave the best performance classifies; as the error rate was very small compared to others.