AGGREGATION STATISTICS AND THE METHODS OF OPERATIONAL RESEARCH FOR WEIGHTING CRITERIA IN MULTIPLE CRITERIA DECISION MAKING

The problem of determining weights of criteria appears in many contexts in multiple criteria decision making. For solving this task it is necessary to determine individual influence of each considered criterion. Each criterion can be either quantitative - when each of attributes can take different values or qualitative - in which case only ranking of weights of criterias is possible. The weights for each criterion could be provided by experiments, tests or as experts opinions. This paper presents methodology for obtaining an aggregated measure, in the form of arithmetic means, for quantitative decision making. The measure is obtained experimentally by combining statistical methodology - based on regression and factor analysis on one hand and operational research approaches - data envelopment analysis as an objective and analytical hierarchical process as subjective method. For practical demonstration of proposed method authors used data on the students' marks, from the official database at the Faculty of organizational sciences, University of Belgrade, to determine the weights of the first-year subjects in order to predict future students' success on each of the study programs, separately.