Muhammad Ilias Amin

2012 Special Issue: Orthogonal least squares based complex-valued functional link network

Functional link networks are single-layered neural networks that impose nonlinearity in the input layer using nonlinear functions of the original input variables. In this paper, we present a fully complex-valued functional link network (CFLN) with multivariate polynomials as the nonlinear functions. Unlike multilayer neural networks, the CFLN is free from local minima problem, and it offers very fast learning of parameters because of its linear structure. Polynomial based CFLN does not require an activation function which is a major concern in the complex-valued neural networks. However, it is important to select a smaller subset of polynomial terms (monomials) for faster and better performance since the number of all possible monomials may be quite large. Here, we use the orthogonal least squares (OLS) method in a constructive fashion (starting from lower degree to higher) for the selection of a parsimonious subset of monomials. It is argued here that computing CFLN in purely complex domain is advantageous than in double-dimensional real domain, in terms of number of connection parameters, faster design, and possibly generalization performance. Simulation results on a function approximation, wind prediction with real-world data, and a nonlinear channel equalization problem exhibit that the OLS based CFLN yields very simple structure having favorable performance.

Complex-valued functional link network design by orthogonal least squares method for function approximation problems

This paper presents a fully complex-valued functional link network (CFLN). The CFLN is a single-layered neural network, which introduces nonlinearity in the input layer using nonlinear functions of the original input variables. In this study, we consider multivariate polynomials as the nonlinear functions. Unlike multilayer neural networks, the CFLN is free from local minima problem, and it offers very fast learning in parameters because of its linear structure. In the complex domain, polynomial based CFLN has an additional advantage of not requiring activation functions, which is a major concern in the complex-valued neural networks. However, it is important to select a smaller subset of polynomial terms (monomials) for faster and better performance, since the number of all possible monomials may be quite large. In this paper, we use the orthogonal least squares method in a constructive fashion (starting from lower degree to higher) for the selection of a parsimonious subset of monomials. Simulation results demonstrate that computing CFLN in purely complex domain is advantageous than in double-dimensional real domain, in terms of number of connection parameters, faster design, and possibly generalization performance. Moreover, our proposed CFLN compares favorably with several other multilayer networks in the complex domain.

Link Prediction in Scientists Collaboration with Author Name and Affiliation

Link prediction task is to discover the indirect relationships from the network based on present information. We studied scientists/authors network and added affiliation information into the authors network in order to enhance the performance of link prediction. In this study, for ranking the collaborations/edges, we have used edge based scoring algorithms rather than node based scoring algorithms and used bipartite graph data structure for finding active authors. We have introduced a feature named Authors Diversity/Similarity score that describes the probability of an author to write with the same or different affiliation. The performance of our proposed algorithm has increased significantly comparing with the existing system.

An Intelligent Decision Support System for Arsenic Mitigation in Bangladesh

Arsenic contamination of groundwater in many nations including Bangladesh shows that this is a global problem. Because of the delayed health effects, poor reporting, and low levels of awareness in some communities, the extent of the adverse health problems caused by arsenic in drinking water is at alarming level in Bangladesh. Also, allocating resources such as tube wells efficiently and effectively to mitigate arsenic hazard is a challenging task in Bangladesh. To allocate resources based on different arsenic hazard parameters, we have developed a Decision Support System that enables the user to observe the effect of allocation policy both in tabular and spatial format using statistical models. We have also developed an algorithm for optimal allocation of resources. A Smart User Interface is designed for the users so that they will find an interactive, user-friendly, intelligible, logical, clear, and sound environment to work with. Finally, we have analyzed and demonstrated the efficacy of our algorithm graphically.