Portfolio management involves modeling risk-return relationships. However, the diverse factors impacting financial markets introduce uncertainty into future portfolio selection. The aim of this study is to propose a portfolio selection model to assist investors in creating the most suitable investment plan in the financial market uncertainty. In this context, a preliminary reduction step is applied to the stocks using the Gustafson-Kessel (GK) algorithm, a fuzzy clustering method, to select portfolio stocks. Later, trapezoidal fuzzy numbers (TrFNs) were defined instead of triangular fuzzy numbers (TFNs) used in the Constrained Fuzzy Analytic Hierarchy Process (AHP) for portfolio selection problems. By using new fuzzy numbers, the weights of the criteria were obtained as TrFNs. Then, a linear programming problem was modeled using the weights of the obtained criteria as a TrFN. For this purpose, a method available in the literature was used that uses price variables in the objective function as TFNs. In this study, a linear programming model that uses these variables as TrFNs is proposed as an alternative to the method that uses the price variables in the objective function as TFNs. In this proposed model, the weights obtained from the Constrained Fuzzy AHP using TrFNs are used as price variables in the objective function of the created linear programming problem. Proposed model then applied to the 48-month return data set of stocks in the Istanbul Stock Exchange 100 (ISE-100) index to determine which stocks the investor should choose and the investment rates investor should make in these stocks. In addition, in order to examine the effectiveness of the proposed model within the scope of the study, portfolio distributions were obtained with different portfolio optimization methods using the same data set and the results were compared.
31 December 2019 in Mehmet Akif Ersoy Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi
In today's developing financial markets, various complex techniques are used in the creation of portfolios that will provide the best return to the investors. In this study, a portfolio selection model that includes investment data and expert opinions is proposed. This model consists of two stages. In the first stage, the weight of the criteria in the portfolio selection problem was determined by the Constrained Fuzzy Analytic Hierarchy Process method proposed by Enea and Piazza. In the second stage, the model proposed by Lai and Hwang was used to solve the problem of fuzzy linear programming to be formed by using the determined criteria weights. These two methods in the literature use triangular fuzzy numbers to solve the problem. The methods used in this study were developed for trapezoidal fuzzy numbers (TrFNs) and an alternative method for portfolio selection problems was proposed.
18 January 2026 in International Journal of Computational Intelligence Systems
Regression analysis estimates model parameters to describe the relationship between a dependent variable and one or more independent variables. While classical regression assumes homogeneity among observations, real-world data often exhibit heterogeneous structures. In such cases, fuzzy logic–based approaches offer an alternative for estimating the unknown parameters of regression models. However, effective results for parameter estimation cannot be obtained for ellipsoidal scattering data in these approaches either. In this study, a method based on the Gustafson-Kessel (GK) clustering algorithm has been proposed to overcome these disadvantages. Since the proposed method aims to perform parameter estimation especially on ellipsoidal data scatters, Davies-Bouldin Index (DBI), one of the clustering validity indices that provide effective results in determining the optimal number of clusters, has been adapted using the Mahalanobis distance. GK clustering algorithm was applied to obtain the membership degrees that determine the contributions of the observations in the identified clusters to the regression model to be constructed. To evaluate the performance of the proposed algorithm, applications were conducted on various datasets, and the results obtained were compared with estimates based on the Fuzzy C-Means (FCM) and Type-2 FCM clustering methods.
Regression analysis is the estimation of model parameters using various statistical methods to model the relationship between the dependent variable and one or more independent variables. In classical regression, although all observations in the data set are assumed to have a homogeneous structure, real-life data often have a heterogeneous structure. In these cases where classical methods are inadequate, fuzzy logic-based approaches are used as an alternative to classical methods for estimating the unknown parameters of the regression model. The first step in the process of estimating parameters is to determine the clusters in the data set and to calculate the membership degrees that express the degree of belonging of each observation to these clusters. These membership degrees will be used as weights in the estimation of model parameters of the relevant observations. In this study, since the aim is to perform parameter estimation specifically on ellipsoidal data distributions, the Davies-Bouldin Index (DBI), one of the clustering validity indices that provides effective results in determining the optimal number of clusters, was used, adapted with the Mahalanobis distance. After determining the optimum number of clusters by DBI, the Gustafson-Kessel (GK) clustering algorithm was applied to obtain the membership degrees of the observations. These membership degrees were evaluated as weights that will contribute to the regression model and models were created. To evaluate the performance of the proposed algorithm, applications were made on various data sets and the obtained results were compared with the estimates based on the Fuzzy C-Means (FCM) clustering method.