Online group streaming feature selection can make full use of the original group structure information in the feature stream to handle the feature selection problem in an online manner.However,most of the existing methods cannot handle the data with ambiguity and uncertainty.To this end,an online group streaming feature selection algorithm based on fuzzy neighbor-hood discrimination index was proposed.A fuzzy neighborhood discrimination index was designed to describe the discriminant information of fuzzy neighborhood granules and extend the related uncertainty measures.On this basis,two strategies,intra-group feature selection and inter-group feature selection,were used to select features with strong approximation ability and non-redundancy.The comparative experiments on eight public datasets verify that the algorithm has better and stable classification performance.
维普
万方数据