A hybrid wrapper-filter approach to detect the source(s) of out-of-control signals in multivariate manufacturing process
- Authors: Huda, Shamsul , Abdollahian, Mali , Mammadov, Musa , Yearwood, John , Ahmed, Shafiq , Sultan, Ibrahim
- Date: 2014
- Type: Text , Journal article
- Relation: European Journal of Operational Research Vol. 237, no. 3 (2014), p. 857-870
- Full Text: false
- Reviewed:
- Description: With modern data-Acquisition equipment and on-line computers used during production, it is now common to monitor several correlated quality characteristics simultaneously in multivariate processes. Multivariate control charts (MCC) are important tools for monitoring multivariate processes. One difficulty encountered with multivariate control charts is the identification of the variable or group of variables that cause an out-of-control signal. Expert knowledge either in combination with wrapper-based supervised classifier or a pre-filter with wrapper are the standard approaches to detect the sources of out-of-control signal. However gathering expert knowledge in source identification is costly and may introduce human error. Individual univariate control charts (UCC) and decomposition of T2 statistics are also used in many cases simultaneously to identify the sources, but these either ignore the correlations between the sources or may take more time with the increase of dimensions. The aim of this paper is to develop a source identification approach that does not need any expert-knowledge and can detect out-of-control signal in less computational complexity. We propose, a hybrid wrapper-filter based source identification approach that hybridizes a Mutual Information (MI) based Maximum Relevance (MR) filter ranking heuristic with an Artificial Neural Network (ANN) based wrapper. The Artificial Neural Network Input Gain Measurement Approximation (ANNIGMA) has been combined with MR (MR-ANNIGMA) to utilize the knowledge about the intrinsic pattern of the quality characteristics computed by the filter for directing the wrapper search process. To compute optimal ANNIGMA score, we also propose a Global MR-ANNIGMA using non-functional relationship between variables which is independent of the derivative of the objective function and has a potential to overcome the local optimization problem of ANN training. The novelty of the proposed approaches is that they combine the advantages of both filter and wrapper approaches and do not require any expert knowledge about the sources of the out-of-control signals. Heuristic score based subset generation process also reduces the search space into polynomial growth which in turns reduces computational time. The proposed approaches were tested by exhaustive experiments using both simulated and real manufacturing data and compared to existing methods including independent filter, wrapper and Multivariate EWMA (MEWMA) methods. The results indicate that the proposed approaches can identify the sources of out-of-control signals more accurately than existing approaches. © 2014 Elsevier B.V. All rights reserved.
Application of optimisation-based data mining techniques to tobacco control dataset
- Authors: Dzalilov, Zari , Zhang, J , Bagirov, Adil , Mammadov, Musa
- Date: 2010
- Type: Text , Journal article
- Relation: International Journal of Lean Thinking Vol. 1, no. 1 (2010), p. 27-41
- Full Text: false
- Reviewed:
- Description: Tobacco smoking is one of the leading causes of death around the world. Consequently, control of tobacco use is an important global public health issue. Tobacco control may be aided by development of theoretical and methodological frameworks for describing and understanding complex tobacco control systems. Linear regression and logistic regression are currently very popular statistical techniques for modeling and analyzing complex data in tobacco control systems. However, in tobacco markets, numerous interrelated factors nontrivially interact with tobacco control policies, such that policies and control outcomes are nonlinearly related.
A new filled function method for nonlinear equations
- Authors: Lin, Yongjian Jian , Yang, Y. , Mammadov, Musa
- Date: 2009
- Type: Text , Journal article
- Relation: Applied Mathematics and Computation Vol. , no. (2009), p.
- Full Text: false
- Description: In this paper, a new global optimization approach based on the filled function method is proposed for solving box-constrained systems of nonlinear equations. We first convert the nonlinear system into an equivalent global optimization problem, and then propose a new filled function method to solve the converted global optimization problem. Several numerical examples are presented and solved by using different local minimization methods, which illustrate the efficiency of the present approach. © 2009 Elsevier Inc. All rights reserved.
Facility location via continuous optimization with discontinuous objective functions
- Authors: Ugon, Julien , Kouhbor, Shahnaz , Mammadov, Musa , Rubinov, Alex , Kruger, Alexander
- Date: 2007
- Type: Text , Journal article
- Relation: ANZIAM Journal Vol. 48, no. 3 (2007), p. 315-325
- Full Text:
- Reviewed:
- Description: Facility location problems are one of the most common applications of optimization methods. Continuous formulations are usually more accurate, but often result in complex problems that cannot be solved using traditional optimization methods. This paper examines the use of a global optimization method - AGOP - for solving location problems where the objective function is discontinuous. This approach is motivated by a real-world application in wireless networks design. © Australian Mathematical Society 2007.
- Description: 2003004859
Quantification of intermarket influence based on the global optimization and its application for stock market prediction
- Authors: Tilakaratne, Chandima , Mammadov, Musa , Hurst, Cameron
- Date: 2006
- Type: Text , Conference paper
- Relation: Paper presented at Integrating AI and Data Mining, 1st International Workshop Proceedings, Hobart, Tasmania : 4th - 5th December, 2006
- Full Text: false
- Reviewed:
- Description: This study investigates how intermarket influences can be used to help the prediction of the direction (up or down) of the next day's close price of the Australian All Ordinary Index (AORD). First, intermarket influences from the potential influential markets on the AORD are quantified by assigning weights for all influential markets. The weights were defined as a solution to an optimization problem which aims to maximise rank correlation between the current day's relative return of the AORD and the weighted sum of lagged relative returns of the potential influential markets. Then, the next day's relative return of the AORD is predicted by applying the neural networks as a classifier. Two different scenarios were compared: 1) using the current day's relative returns of different sets of influential markets as separate inputs; and, 2) using only the weighted sum of these relative returns as a "combined market". The results revealed that the second approach provides better predictions in all cases. This shows the effectiveness of the proposed approach for quantifying intermarket influences and the potential of using the "weighted combined markets" for the prediction
- Description: E1
- Description: 2003001609
An optimization approach to identifying drugs responsible for adverse drug reactions
- Authors: Mammadov, Musa , Banerjee, Arunava
- Date: 2005
- Type: Text , Conference paper
- Relation: Paper pesented at Sixteenth Australasian Workshop on Combinatorial Algorithms, AWOCA 2005, Ballarat, Victoria : 18th-21st September 2005 p. 185-200
- Full Text: false
- Description: In this paper we develop an optimization approach for the study of Adverse Drug Reaction (ADR) problems. This approach is based on drug-reaction relationships represented in the form of a vector weights, which can be defined as a solution to some global optimization problem. Although it can be used for solving many ADR problems, we concentrate on the problem of accurate identification of drugs that are responsible for reactions that have occurred. Based on drug-reaction relationships, we formulate this problem as an optimization problem. The approach is applied to Australian Adverse Drug Reaction Advisory Committee (ADRAC) database. We take a comprehensive approach to considering all reaction classes which combines 18 SOC (System Organ Class), as well as the sub-classes of reaction classes Blood, Body, Neurological and Cardiovascular. The numerical experiments provided high accuracy in prediction of suspected drugs reported in ADRAC database.
- Description: 2003001383
An optimization approach to the study of drug-drug interactions
- Authors: Mammadov, Musa , Banerjee, Arunava
- Date: 2005
- Type: Text , Conference paper
- Relation: Paper pesented at Sixteenth Australasian Workshop on Combinatorial Algorithms, AWOCA 2005, Ballarat, Victoria : 18th-21st September 2005 p. 201-216
- Full Text:
- Description: Drug-drug interaction is one of the important problems of Adverse Drug Reaction (ADR). In this paper we develop an optimization approach for the study of this problem. This approach is based on drug-reaction relationships represented in the form of a vector of weights, which can be defined as a solution to some global optimization problem. Although this approach can be used for solving many ADR problems, we concentrate here only on drug-drug interactions. Based on drug-reaction relationships, we formulate this problem as an optimization problem. The approach is applied to different classes of reactions from the Australian Adverse Drug Reaction Advisory Committee (ADRAC) database.
- Description: 2003001384
H-infinity via a nonsmooth, nonconvex optimization approach
- Authors: Mammadov, Musa , Orsi, Robert
- Date: 2005
- Type: Text , Journal article
- Relation: Pacific Journal of Optimization Vol. 1, no. 2 (2005), p. 405-420
- Full Text: false
- Reviewed:
- Description: A numerical method for solving the H-infinity synthesis problem is presented. The problem is posed as an unconstrained, nonsmooth, nonconvex minimization problem. The optimization variables consist solely of the entries of the output feedback matrix. No additional variables, such as Lyapunov variables, need to be introduced. The main part of the optimization procedure uses a line search mechanism where the descent direction is defined by a recently introduced dynamical systems approach. Numerical results for various benchmark problems are included in the paper. In addition, the effectiveness of a preliminary part of the algorithm for successfully and quickly finding stabilizing controllers is also demonstrated.
- Description: C1
- Description: 2003001382