We report a benchmarking of neural networks and regression techniques in a time series forecasting task. The estimation errors, computing costs and additional information obtained by Bayesian neural networks are compared with other neural network models and with Multivariate Adaptive Regression Splines (MARS). The Mackey Glass time series in chaotic regime was used to generate the two data sets...

The solution of linear inverse problems obtained by means of regularization theory has the structure of a neural network similar to classical RBF networks. However, the basis functions depend in a nontrivial way on the specific linear operator to be inverted and the adopted regularization strategy. By resorting to the Bayesian interpretation of regularization, we show that such networks can be ...

This study was conducted to rate the land characteristics of corn in hot areas based on artificial neural networks and regression models. For this purpose, 63 corn fields were selected in southern Iran. In each farm, a pedon was excavated, described and sampled. A questionnaire was completed for each farm. A stepwise regression model was used to study the relationship between land characteristi...

one of the most important issues always facing banks and financial institutes is the issue of credit risk or the possibility of failure in the fulfillment of obligations by applicants who are receiving credit facilities. the considerable number of banks’ delayed loan payments all around the world shows the importance of this issue and the necessary consideration of this topic. accordingly, many...

Generally, learning is performed so as to minimize the sum of squared errors between network outputs and training data. Unfortunately, this procedure does not necessarily give us a network with good generalization ability when the number of connection weights are relatively large. In such situation, overfitting to the training data occurs. To overcome this problem, there are several approaches ...

We propose a scalable Gaussian process model for regression by applying a deep neural network as the feature-mapping function. We first pre-train the deep neural network with a stacked denoising auto-encoder in an unsupervised way. Then, we perform a Bayesian linear regression on the top layer of the pre-trained deep network. The resulting model, Deep-Neural-Network-based Gaussian Process (DNN-...

Deep neural networks (DNNs) often require good regularizers to generalize well. Currently, state-of-the-art DNN regularization techniques consist in randomly dropping units and/or connections on each iteration of the training algorithm. Dropout and DropConnect are characteristic examples of such regularizers, that are widely popular among practitioners. However, a drawback of such approaches co...

The emulation of mechanical systems is a popular application of artificial neural networks in engineering. This paper examines general principles of modelling mechanical systems by feedforward artificial neural networks (FFANNs). The slow convergence issue associated with the highly parallel and redundant structure of FFANN systems is addressed by formulating criteria for constraining network p...