<ul><li>RelChaNet is a supervised feature selection algorithm that utilizes neuron pruning and regrowth in a dense neural network's input layer.</li><li>For pruning, a relative change metric is used to measure the impact a feature has on the network.</li><li>In addition, an extension is proposed to dynamically adapt the size of the input layer during runtime.</li><li>Experimental results on 13 datasets demonstrate that RelChaNet outperforms existing methods, with a 2% increase in accuracy on the MNIST dataset.</li></ul>

RelChaNet: Neural Network Feature Selection using Relative Change Scores

Discover more