Efficiency of New Feature Selection Method Based on Neural Network

Nicole Challita (1) , Mohamad Khalil (2) , Pierre Beauseroy (1)

1 LM2S - Laboratoire Modélisation et Sûreté des Systèmes
2 LU - Lebanese University [Beirut]

In order to monitor a system, the number of measurements and features gathered can be huge. But it is desirable to keep only the important features to reduce the processing demand. The problem is therefore to select a subset of features to obtain the best possible classification performance. In this purpose, many feature selection algorithms have been proposed. In a previous work, we have proposed a new feature selection method inspired by neural network and machine learning. This new method selects the best features using sparse weights of the input features in the neural network. In this paper, we study the performance of this method on simulated data.

Domains: Engineering Sciences [physics]

Nicole Challita, Mohamad Khalil, Pierre Beauseroy. Efficiency of New Feature Selection Method Based on Neural Network. 2018 International Conference on Computer and Applications (ICCA), Aug 2018, Beirut, Lebanon. pp.1-320, ⟨10.1109/COMAPP.2018.8460292⟩. ⟨hal-03320925⟩