The popularity of feed-forward neural networks such as multilayer perceptrons and radial basis function networks is to a large extent due to their universal approximation capability. This paper concerns its theoretical principles, together with the influence of network architecture and of the distribution of training data on this capability. Then, the possibility to exploit this influence in order to improve the approximation capability of multilayer perceptrons by means of cross-validation and boosting is explained. Although in theory, the impact of both methods on the approximation capability of feed-forward networks is known, they are still not common in real-world applications. Therefore, the paper documents usefulness of both methods on a detailed case study in materials science.
The N250r is a face-sensitive event-related potential (ERP) deflection whose long-term memory sensitivity remains uncertain. We investigated the possibility that long-term memory-related voltage changes are represented in the early ERP's to faces but methodological considerations could affect how these changes appear to be manifested. We examined the effects of two peak analysis procedures in the assessment of the memory-sensitivity of the N250r elicited in an old/new recognition paradigm using analysis of variance (ANOVA) and artificial neural networks (ANN's). When latency was kept constant within subjects, ANOVA was unable to detect differences between ERP's to remembered and new faces; however, an ANN was. Network interpretation suggested that the ANN was detecting amplitude differences at occipitotemporal and frontocentral sites corresponding to the N250r. When peak latency was taken into account, ANOVA detected a significant decrease in onset latency of the N250r to remembered faces and amplitude differences were not detectable, even with an ANN. Results suggest that the N250r is sensitive to long-term memory. This effect may be a priming phenomenon that is attenuated at long lags between faces. Choice of peak analysis procedures is critical to the interpretation of phasic memory effects in ERP data.
Correct detection of input and output parameters of a welding process is significant for successful development of an automated welding operation. In welding process literature, we observe that output parameters are predicted according to given input parameters. As a new approach to previous efforts, this paper presents a new modeling approach on prediction and classification of welding parameters. 3 different models are developed on a critical welding process based on Artificial Neural Networks (ANNs) which are (i) Output parameter prediction, (ii) Input parameter prediction (reverse application of output prediction model) and (iii) Classification of products. In this study, firstly we use Pareto Analysis for determining uncontrollable input parameters of the welding process based on expert views. With the help of these analysis, 9 uncontrollable parameters are determined among 22 potential parameters. Then, the welding process of ammunition is modeled as a multi-input multi-output process with 9 input and 3 output parameters. 1st model predicts the values of output parameters according to given input values. 2nd model predicts the values of correct input parameter combination for a defect-free weld operation and 3rd model is used to classify the products whether defected or defect-free. 3rd model is also used for validation of results obtained by 1st and 2nd models. A high level of performance is attained by all the methods tested in this study. In addition, the product is a strategic ammunition in the armed forces inventory which is manufactured in a limited number of countries in the world. Before application of this study, the welding process of the product could not be carried out in a systematic way. The process was conducted by trialand- error approach by changing input parameter values at each operation. This caused a lot of costs. With the help of this study, best parameter combination is found, tested, validated with ANNs and operation costs are minimized by 30%.