Random KNN consists of an ensemble of base k-nearest neighbor models, each constructed from a random subset of the input variables. Random KNN can be used to select important features using the RKNN-FS algorithm. RKNN-FS is an innovative feature selection procedure for ‘small n, large p problems.’ Random KNN (no bootstrapping) is fast and stable compared with Random Forests. The rknn R package implements Random KNN classification, regression and variable selection algorithms.
• KNN is stable, no hierarchical structure
• Final model can be a single KNN (vs. many trees)
• Local method: robust for complex data structure
• Automatically re-train, incremental learning
• Easy to implement … Random KNN (RKNN)