A KNN Algorithm Based on Mixed Normalization Factors

K-Nearest Neighbor (KNN) algorithm is a non-parametric statistical method used for classification and regression and it is very simple and effective. KNN may require a large amount of memory or space to store all data, and using distance or proximity measurement methods may crash at very high dimensions (with many input variables), which may have a negative impact on the performance of the algorithm on your problem. However, the performance of KNN will decrease for the large-scale datasets and high-dimensional data. So, we design a mixed normalization factors based on different data sets. Based on the least mean square algorithm (LMS), we design the normalization factors by min-max normalization and mean normalization. Comparison of other good classification, the proposed KNN algorithm based on mixed normalization factors has best performance.