Research on small files classification based on improved KNN algorithm and pretreatment strategy

This article which combines MapReduce model with mass data processing innovatively, proposes small files classification and pretreatment strategy research on mass data. The described method provides more convenience for the parallel computing characteristics of MapReduce architecture, and saves a large amount of processing time. Meanwhile, the classification method is proved to be efficient and reliable through some experiments. The strategy of the paper can be widely applied to document classification and clustering research and application. © 2015, ICIC International.