Neural network software

Abstract: From the Publisher: When applied to the world of finance, neural networks are automated trading systems, based on mapping inputs and outputs for forecasting probable future values. In Neural Networks for Financial Forecasting - the first book to focus on the role of neural networks specifically in price forecasting - traders are provided with a solid foundation that explains how neural nets work, what they can accomplish, and how to construct, use, and apply them for maximum profit. It is written by an acknowledged authority who is, himself, the developer of several successful networks. Neural Networks for Financial Forecasting enables you to develop a usable, state-of-the-art network from scratch all the way through completion of training. There are spreadsheets and graphs throughout to illustrate key points, and an appendix of valuable information, including neural network software suppliers and related publications.

Abstract: This paper compares empirically the predictive performance of two different methods of software reliability prediction: 'neural networks' and 'recalibration for parametric models'. Both methods were claimed to predict as good or better than the conventional parametric models that have been used-with limited results so far. Each method applied its own predictability measure, impeding a direct comparison. To be able to compare, this study uses a common predictability measure and common data-sets. This study reveals that neural networks are not only much simpler to use than the recalibration method, but that they are equal or better trend (variable term) predictors. The neural network prediction is further improved by preparing the data with a running average, instead of the traditionally used averages of grouped data points. Neural network predictions do not depend on prior known models. Off-the-shelf neural network software tools make it easy to apply the method.

Abstract: The present paper illustrates the feasibility of using neural network models for priority assessment of highway pavement maintenance needs. Since neural networks are developed to mimic the decision-making process of human beings and do not require users to predefine a mathematical equation relating pavement conditions to priority ratings, they offer an attractive means by which the priority setting process by highway maintenance personnel can be simulated. In the present study, the ability of a simple back-propagation neural network was tested separately with three different priority-setting schemes, using a general-purpose microcomputer-based neural network software. The priority-setting schemes include a linear function relating priority ratings to pavement conditions, a nonlinear function, and subjective priority assessments obtained from a pavement engineer. For the first two schemes, noise was also introduced to examine how it would affect the performance of the neural network. Test results are positive and indicative of the potential of neural networks as a useful tool that highway agencies can use for priority rating in maintenance planning at the network level.

Abstract: We study in this paper several methods that allow one to use interval data as inputs for Multi-layer Perceptrons. We show that interesting results can be obtained by using together two methods: the extremal values method which is based on a complete description of intervals, and the simulation method which is based on a probabilistic understanding of intervals. Both methods can be easily implemented on top of existing neural network software.