1. The Graeco-Latin Square and Hyper Graeco-Latin Square Designs

    W.H. Moolman

    Walter Sisulu University, Mthatha, South Africa.

    As an experimental design model the Graeco-Latin square is an extension of a Latin square and can simultaneously control three sources of nuisance variability. The following aspects of this model will be discussed: a brief history, estimation and ANOVA, use for the analysis of experimental data (example with R code given), model generation and a test for non-additivity. An R example of the Hyper Graeco-Latin square model, which extends the Graeco-Latin square to controlling four sources of nuisance variability, will also be discussed.

    Keywords: Graeco-Latin, Euler, mutually orthogonal, non-additivity, design generation, Youden square, Hyper Graeco-Latin, Replications.

    Pages: 211 – 220 | Full PDF Paper
  2. Estimation of Users Satisfaction in Electronic Banking: Neural Network Approach

    Stefan Zdravković1, Jelena Peković2 , Aleksandar Jovanović3

    1. Faculty of Economics, University of Kragujevac, Djure Pucara Starog 3, 34000, Kragujevac, Serbia.

    2. Faculty of Economics, University of Kragujevac, Djure Pucara Starog 3, 34000 Kragujevac, Serbia.
    3. Faculty of Engineering, University of Kragujevac, Sestre Janjic 6, 34000, Kragujevac, Serbia.

    The rapid progress of information technologies has resulted in the development of e-banking. The significant range in consumer demands and lifestyles have imposed the need for banks to use new technologies to communicate with clients. The main advantage of e-banking is the ability to conduct transactions at any time and anywhere. For banks, as providers of e-banking services, it is important to identify factors that influence customer satisfaction. In this paper, we deal with users attitudes that were examined using a pre-prepared survey, which is composed of relevant issues taken from the literature in order to assess the user’s satisfaction. The results thus obtained relate to the subjective level of satisfaction of each user. Four parameters were selected that can affect users, and for each of them, the degree of influence on the ultimate customer satisfaction was examined. After then we develop a neural network that approximates the function of an objective level of user satisfaction, taking into account the views of all previous users and matching them. For the last 5 users, who completed the survey, both subjective and objective level of satisfaction was presented. This method for estimating the objective level of customer satisfaction can serve banks as a parameter in estimating the effects of the applied e-banking system. Data processing was performed using SPSS statistical software. Neural network was developed using MATLAB statistical software.

    E-banking, Users satisfaction, Neural network.

    Pages: 221 – 232 | Full PDF Paper