International Journal of Information Technology & Decision Making

A new multivariate statistical technique is obtained for comparing and combining two or more data sets each of which has a different number of respondents but the same variables. This approach can be considered as dual to such techniques as partial least squares, also known as inter-battery factor analysis and robust canonical correlation analysis for two data sets. It is shown that the problem can be reduced to the eigenproblem of the product of correlation matrices of each data set. The technique is generalized to three or more data sets in an eigenproblem of block-matrices of the correlations within each data set. This type of multivariate analysis can serve various practical problems of integration of data obtained from heterogeneous sources, particularly, for data merging in constructing data warehouses.

Data envelopment analysis (DEA) is a nonparametric linear programming technique for measuring the relative efficiency of decision making units (DMUs) on the basis of multiple inputs and outputs. DEA assessments, however, are proved to be sensitive to extreme units that deviate substantially in their input/output patterns. In this paper we introduce an approach for handling extreme observations in DEA, i.e., observations that exhibit irregularly high values in some outputs and/or low values in some inputs. Unlike the usual practice of removing such observations, we retain them in the production possibility set reducing their impact on the other units. Our modeling approach is based on the concept of diminishing returns, assuming that the contribution of an output (input) to the efficiency score diminishes as the output increases beyond a pre-specified level, i.e., the level beyond which a value is characterized as extreme. According to our approach the original data set is transformed to an augmented data set, where standard DEA models can then be applied, remaining thus in the grounds of the standard DEA methodology. We illustrate our approach with a numerical example.

The fact that the relationships among the returns of financial assets tend to be nonlinear and time-varying has important implications for asset allocation. To describe these two features, this paper first combines a copula function with the Markov switching technique to model the dependence structure across assets and then builds on this Markov Switching Copula model to present a procedure for the timing of portfolio adjustments. Our empirical evidence confirms that the dependence structure between high-risk and low-risk stocks in the Shanghai and Shenzhen markets is not static but switches between regimes over the course of the sample horizon considered in this paper. More importantly, as a result of such regime-switching characteristics of their dependence structure, our analysis of the out-of-sample asset allocation performance indicates that employing the procedure proposed in this paper to identify regime changes and decide when to adjust portfolio weights allows investors with the Constant Relative Risk Aversion utility to achieve both higher realized returns and higher certainty equivalent rate of returns than does the use of strategies based on static models.

Business simulators are powerful tools for both supporting the decision-making process of business managers as well as for business education. An example is SIMBA (SIMulator for Business Administration), a powerful simulator which is currently used as a web-based platform for business education in different institutions. In this paper, we propose the application of reinforcement learning (RL) for the creation of intelligent agents that can manage virtual companies in SIMBA. This application is not trivial, given the particular intrinsic characteristics of SIMBA: it is a generalized domain where hundreds of parameters modify the domain behavior; it is a multi-agent domain where both cooperation and competition among different agents can coexist; it is required to set dozens of continuous decision variables for a given business decision, which is made only after the study of hundreds of continuous variables. We will demonstrate empirically that all these challenges can be overcome through the use of RL, showing results for different learning scenarios.

Price promotion and product recommendation are important tactics to gain market share in the e-commerce context. To increase sales and enhance promotion profits, practitioners often recommend non-discounted products in the promotional campaign for a product. However, no analytical model is hitherto available to jointly optimize the price discount for the promoted product and the product portfolio to recommend. This paper provides a probability model to complete the task. The proposed model motivates customers through an attractive price discount for the promoted product, while simultaneously encouraging customers to purchase the non-discounted products through the recommendation system. The numerical studies show that the proposed method attains higher profits than do conventional methods. Finally, we offer managerial insights and provide useful guidelines to help e-tailers to make the most profitable online promotional decisions.

In many countries, including Iran, Provincial Departments of Physical Education try to develop the athletic sports and sports for all in their related areas (state), using the government resources. Their success rate has always been an important subject for the top sports managers of country. In this paper we use data envelopment analysis (DEA) and analytic hierarchy process (AHP) techniques for analyzing the performance of physical education organizations in Iran. Some convex and nonconvex DEA models have been used. Afterwards, we have used the Shannon's entropy for aggregating the results obtained from different models and providing a final efficiency score (FES) and a unified ranking. It can be seen that, in the ranking approach provided in this paper the most productive scale size (MPSS) units have the best rank (see Proposition 1). Our findings reveal that the average of FESs of the states is 0.472635 and 50% of the states have an FES more than this average. Classifying the sates to five efficiency classes, "Excellent, Good, Middle, Weak and Very Weak", the percentage of the states belonging to these classes are 6.7, 30, 16.6, 36.7 and 10, respectively. Also, some correlation and difference studies have been carried out using the Pearson's correlation and student's t-tests. Finally, comparisons between the results of some relevant existing publications and those given in the present paper are addressed.

Utility mining has recently been discussed in the field of data mining. A utility itemset considers both profits and quantities of items in transactions, and thus its utility value increases with increasing itemset length. To reveal a better utility effect, an average-utility measure, which is the total utility of an itemset divided by its itemset length, is proposed. However, existing approaches use the traditional average-utility upper-bound model to find high average-utility itemsets, and thus generate a large number of unpromising candidates in the mining process. The present study proposes an improved upper-bound approach that uses the prefix concept to create tighter upper bounds of average-utility values for itemsets, thus reducing the number of unpromising itemsets for mining. Results from experiments on two real databases show that the proposed algorithm outperforms other mining algorithms under various parameter settings.