Optimal Architecture Deep Neural Network for Regression

These codes perform optimal architecture deep neural network model generation for regression problems and illustrate e-nose application.

Deep Neural Network (DNN) models require a suitable neural architecture (topology) for data driven modeling. Avoiding overfitting or underfitting problems and achieving a satisfactory generalization are main concerns in modeling tasks. To deal with these concerns, optimization of neural architecture according to the dataset is an effective solution.

These codes perform optimization of deep neural network model architecture for regression problems. An example is shown for e-nose application.