Overview on last advances of feature selection
Saâd Nouinou, Abdellatif El Afia, Sanaa El Fkihi · 2018
In the context of learning from high-dimensional datasets, the performance of machine learning algorithms decreases and can even deteriorate in the presence of data noisiness. Feature selection is a dimension reduction process for encountering the curse of dimensionality, which can also solve the noise issue. This paper aims to describe the latest advances in feature selection. The contributions are investigated as attempts for overcoming challenges related to the feasibility, computational complexity, accuracy and reliability. This paper may allow the researcher to take a broad view of developments in feature selection.