Speeding up feature selection: A deep-inspired network pruning algorithm

To address the high-dimensionality of big data, numerous iterative algorithms have been introduced including least absolute shrinkage selection operator (Lasso) and iteratively sure independent screening (ISIS). However, the iterative nature of these algorithms renders the computational cost of retraining the learning model impractical. We take advantage of this key observation to propose a novel non-iterative algorithm inspired by deep neural network input reconstruction. Our proposed technique provides faster feature selection by training the model once. This is achieved by exploiting the cross entropy error between the input and the estimates. Simulation studies support our approach on several real world datasets. For rigor, a comparative analysis of the computational complexity is also provided to assert the advantage of our approach.