Maziar Salahi

Work place: Dept. of Applied Mathematics, University of Guilan, Rasht, Iran



Research Interests: Data Structures and Algorithms, Analysis of Algorithms, Combinatorial Optimization, Randomized Algorithms


Maziar Salahi is an associate professor at the department of Applied Mathematics in University of Guilan, Rasht, Iran. He received his PhD from McMaster University, Canada in 2006 in the field of optimization. His main research focus is on developing efficient algorithms for solving optimization problems.

Author Articles
Supplier Segmentation using Fuzzy Linguistic Preference Relations and Fuzzy Clustering

By Pegah Sagheb Haghighi Mahmoud Moradi Maziar Salahi

DOI:, Pub. Date: 8 Apr. 2014

In an environment characterized by its competitiveness, managing and monitoring relationships with suppliers are of the essence. Supplier management includes supplier segmentation. Existing literature demonstrates that suppliers are mostly segmented by computing their aggregated scores, without taking each supplier’s criterion value into account. The principle aim of this paper is to propose a supplier segmentation method that compares each supplier’s criterion value with exactly the same criterion of other suppliers. The Fuzzy Linguistic Preference Relations (LinPreRa) based Analytic Hierarchy Process (AHP) is first used to find the weight of each criterion. Then, Fuzzy c-means algorithm is employed to cluster suppliers based on their membership degrees. The obtained results show that the proposed method enhances the quality of the previous findings.

[...] Read more.
Modeling Truncated Loss Data of Operational Risk in E-banking

By Maryam Pirouz Maziar Salahi

DOI:, Pub. Date: 8 Nov. 2013

Operational risk is an important risk component for financial institutions, especially in E-banking. Large amount of capital that are assigned to decrease this risk are evidence to this subject. One of the most important factors for modeling operational risk to estimate capital charge is loss data collections of banks. But sometimes for reasons like decreasing the costs, banks save only the losses above some determined thresholds at their database. For achieving accurate capital charge, this threshold should be considered in determining capital charge. This paper focuses on modeling truncated loss data above some given threshold. We discuss several statistical methods for modeling truncated data, and suggest the best one for modeling truncated loss data. We have tested our suggested model for some operational loss data samples. Our results indicate that our approach can be useful for increasing accuracy of estimating operational risk capital charge in E-banking.

[...] Read more.
Other Articles