Showing papers by "Babak Abbasi published in 2008"

Estimating parameters of the three-parameter Weibull distribution using a neural network

Abstract: Weibull distributions play an important role in reliability studies and have many applications in engineering. It normally appears in the statistical scripts as having two parameters, making it easy to estimate its parameters. However, once you go beyond the two parameter distribution, things become complicated. For example, estimating the parameters of a three-parameter Weibull distribution has historically been a complicated and sometimes contentious line of research since classical estimation procedures such as Maximum Likelihood Estimation (MLE) have become almost too complicated to implement. In this paper, we will discuss an approach that takes advantage of Artificial Neural Networks (ANN), which allow us to propose a simple neural network that simultaneously estimates the three parameters. The ANN neural network exploits the concept of the moment method to estimate Weibull parameters using mean, standard deviation, median, skewness and kurtosis. To demonstrate the power of the proposed ANN-based method we conduct an extensive simulation study and compare the results of the proposed method with an MLE and two moment-based methods.

Detection and classification mean-shifts in multi-attribute processes by artificial neural networks

Abstract: To monitor the quality of a multi-attribute process, some issues arise. One of them being the occurrence of a high number of false alarms (type I error) and the other an increase in the probability of not detecting defects when the process is monitored by a set of independent uni-attribute control charts. In this paper, based upon the artificial neural network capabilities we develop a new methodology to overcome this problem. We design a perceptron neural network to monitor either the proportions of several types of product nonconformities (instead of using several np charts) or the number of different types of defects (instead of using several c charts) in a product. Moreover, while the proposed method possesses the ability to be applied for small sample sizes, it is also able to diagnose the mean shift online. We present two simulation experiments in which the proportions of several types of nonconformities are monitored. In addition, we present one more simulation experiment in which the number of different types of defect is controlled. We also compare the performance of the proposed methodology with the ones from the Mnp and T2 charts for multi-attribute processes. The results of the simulation studies are encouraging.

Generating correlation matrices for normal random vectors in NORTA algorithm using artificial neural networks

Abstract: Generating multivariate random vectors is a crucial part of the input analysis involved in discrete-event stochastic simulation modeling of multivariate systems. The NORmal-To-Anything (NORTA) algorithm, in which generating the correlation matrices of normal random vectors is the most important task, is one of the most efficient methods in this area. In this algorithm, we need to solve the so-called correlation-matching problem in which some complicated equations that are defined to obtain the correlation matrix of normal random variables need to be solved. Many researchers have tried to solve these equations by three general approaches of (1) solving nonlinear equations analytically, (2) solving equations numerically, and (3) solving equations by simulation. This paper suggests the use of artificial neural networks, called Perceptron, to solve the corresponding problem. Using three simulation experiments, the applicability of the proposed methodology is described and the results obtained from the proposed method to the ones from solving the equations numerically are compared. The results of the simulation experiments show that the proposed method works well. © 2008 World Academic Press, UK. All rights reserved.