Research

Feature selection is used to select a subset of relevant features in machine learning, and is vital for simplication, improving eciency and reducing overtting. In lter-based feature selection, a statistic such as correlation or entropy is computed between each feature and the target variable to evaluate feature relevance. A relevance threshold is typically used to limit the set of selected features, and features can also be removed based on redundancy (similarity to other features). Some methods are designed for use with a specic statistic or certain types of data. We present a new lter-based method called Relevance-Redundancy Domi- nance that applies to mixed data types, can use a wide variety of statis- tics, and does not require a threshold. In experiments it outperformed published methods on credit data.