Seybold Report ISSN: 1533-9211
Meenakshi, Ramachandra AC
Vol 17, No 11 ( 2022 ) | Licensing: CC 4.0 | Pg no: 2977-2985 | Published on: 30-11-2022
Abstract
In this paper we discuss importance of selecting features and vario¬us approaches for feature selection during Machine Learning modeling. Data cleaning and Feature selection are two important steps carried out in the beginning of every machine learning model designing. While final features selected for modeling have a significant effect on performance, on the other side insignificant features give a negative effect of performance. For Deep Learning models, explicit feature selection is not advisable due to presence of its inbuilt internal feature selection while modelling itself. Depending on domain and dataset, still feature selection can be applicable for reducing time and space complexity. It also helps to explore weak, noisy, irrelevant features present in the dataset.
Keywords:
Deep Learning, Feature Selection, Machine Learning, Supervised, Unsupervised