An Extended Study on High-Level K-Nearest Neighbors (HLKNN): Enhanced Classification Using Second-Level Neighbor Analysis on Benchmark Datasets

This work presents a student-level reproduction and extension of the High-Level K-Nearest Neighbors (HLKNN) algorithm proposed by Al-Hmouz (2023). The study evaluates HLKNN against the standard KNN algorithm using the Iris, Breast Cancer Wisconsin, and UCI Phishing Website Detection datasets. Experimental results demonstrate improved classification accuracy through second-level neighbor analysis. This manuscript was prepared as part of the MCA Machine Learning Internal Assessment at Chanakya University and is shared for educational purposes. It has not undergone peer review.