Several methods of fuzzy rule based classiication are applied to medical data. Features from patient data, collected in a clinic, are pre-processed before being fed into fuzzy neural networks, where fuzzy rule based classiication systems are generated. Additionally, results obtained from feed-forward architectures such as standard-backpropagation networks, radial basis function nets (RBF) and D...

Maintaining software has always been a difficult task. In most cases,maintaining software takes greater time and effort than its development time. Also, predicting the software maintainability makes the process more complex. Reliability in terms of application software is characterized as the functioning of software i.e. carrying out its operations without any failure is the most important requ...

This paper presents an innovative method to evaluate the credibility of network connections based on neural network with back propagation (BP) network model and Levenberg-Marquardt (LM) optimization algorithm. In it, the second-level index values are taken as input and expected evaluation values as output, based on which, L test samples are used to test the trained neural network. At last, a pr...

In this paper we explore how machine learning techniques can be applied to the discovery of efficient mathematical identities. We introduce an attribute grammar framework for representing symbolic expressions. Given a grammar of math operators, we build trees that combine them in different ways, looking for compositions that are analytically equivalent to a target expression but of lower comput...

In this study, the problem of updating a printer characterization in response to systematic changes in print-device characteristics is addressed with two distinct approaches: the creation of corrective models used in conjunction with an existing device model, and the re-evaluation of regression-model parameters using an augmented characterization data set. Several types of corrective models are...

Using a Kalman-Trained Neural Network M. Birgmeier Abstract| This paper presents a neuralnetwork-based equalizer for a digital communication system. In this equalizer, the neural network is adapted to the optimum decision boundaries as de ned by the channel and noise characteristics. This problem has already been considered by Gibson et al. [1], but for more complex decision boundaries, the sim...

Prediction of the performance of international joint ventures remains a relatively underresearched area, yet its importance is well recognized due to the tremendous surge in joint venture activities in the past decade. Data on 1,463 Sino–Hong Kong joint ventures were gathered and those that appeared on the Honor Roll of the China Association of Enterprises with Foreign Investment are identified...

Investigation for the customer behavior is an emerging field that many studies fouces on in past. In of bank, predicting whether customr will leave or not important question. This study mainly tries to figure out what most influencing factor a when they choose exit tone financial institution. It company get rational prediction, which could help them make right decosison maintain and retain cust...

through the last three decades different seismological and engineering approaches for the generation of artificial earthquakes have been proposed. selection of an appropriate method for the generation of applicable artificial earthquake accelerograms (aeas) has been a challenging subject in the time history analysis of the structures in the case of the absence of sufficient recorded accelerogra...

In computer-aided drug discovery, quantitative structure activity relation models are trained to predict biological from chemical structure. Despite the recent success of applying graph neural network this task, important information such as molecular chirality is ignored. To fill crucial gap, we propose Molecular-Kernel Graph NeuralNetwork (MolKGNN) for representation learning, which features ...