Semi-supervised Classification Algorithm Using Fuzzy Nearest Neighborhood Label Propagation

A Fuzzy Nearest Neighborhood Label Propagation algorithm (FNNLP) is proposed,which is sample distribution independent and is simple with only one parameter.Based on the similarity connections of the k nearest data,FNNLP classifies the unlabeled data by making the labels propagate from labeled data to unlabeled ones.Finally,two promising experiment results of synthesized data and UCI dataset classifications verify the simplicity and effectiveness of FNNLP.