A dimensionality reduction based on rough set theory for complex massive data

摘要

Dimensionality reduction is the important topic for data mining and pattern recognition. Many dimensionality reduction methods for complex massive data have been proposed. Due to massive data have many kinds of data such as: noise, inconsistent and incomplete information. The dimensionality reduction task is difficu to date, there are no efficient approaches for dimensionality reduction in complex massive data. Here we attempt to provide a quick approach to deal with this issue. At first, two kinds of efficient attribute measurement methods are presented, and discuss the relationships between two kinds of dimensionality reduction; what's more, two dimensionality reduction methods are designed respectively; Finally, experimental results verify the feasible of the designed algorithms.