k nearest Neighbors classifies or predicts quantitative value for new samples using voting/average values of the nearest neighbors of the analyzed sample. Prior to model training all descriptors are normalized to standard deviation of 1 and mean value of 0.
Algorithm parameters include
- Distance (Euclidian distance, Pearson correlation)
KNN neighbors
The specified number of nearest neighbors will be used, unless this value is 0. In the later case this number will be optimized in range from 1 to Max KNN neighbors.Max KNN neighbors.
Specifies maximum number of nearest neighbors for automatic determination of the KNN neighbors number.