This article concerns the application of the Mixture of Local Expert Models (MLEM) to predict the daily and monthly price of the Sugar No. 14 contract in the New York Board of Trade. This technique can be seen as a forecasting method that performs data exploratory analysis and mathematical modeling simultaneously. Given a set of data points, the basic idea is as follows: 1) a Kohonen Neural Network is used to divide the data into clusters of points, 2) several modeling techniques are then used to construct competing models for each cluster, 3) the best model for each cluster is then selected and called the Local Expert Model. Finally, a so-called Gating Network combines the outputs of all Local Expert Models. For comparison purposes, the same modeling techniques are also evaluated when acting as Global Experts, i. e., when the technique uses the entire data set without any clustering.Keywords: mixture of local expert models; forecasting time series; neural networks. ResumoEste artigo aborda a aplicação de Modelos da Composição de Especialistas Locais (MCEL) para previsão de preços diários e mensais da comodity açúcar da bolsa de valores de Nova York. Esta técnica pode ser vista como método de previsão que realiza simultaneamente análise exploratória de dados assim como modelagem matemática. Dado um conjunto de dados, a idéia básica é a seguinte: 1) uma Rede Neural de Kohonen é utilizada para dividir o conjunto de dados em clusters, 2) várias técnicas de modelagem são utilizadas para calibrar modelos para cada cluster, 3) o melhor modelo para cada cluster é selecionado e denominado Modelo de Especialista Local. Finalmente, uma Rede Supervisora combina as saídas de cada um dos Modelos de Especialista Local. Com a finalidade de comparação, as mesmas técnicas de modelagem são também avaliadas atuando como Especialistas Globais, ou seja, quando as técnicas usam o conjunto de dados único, sem clusters.Palavras-chave: modelos da composição de especialistas locais; previsão de séries temporais; redes neurais.

No abstract

No abstract

No abstract

ResumoNeste trabalho são propostos dois modelos tipo DEA (Data Envelopment Analysis) em que um novo input (ou um novo output) deve ser distribuído de maneira justa para todas as unidades (DMUsDecision Making Units) de uma corporação. Uma característica interessante dos modelos propostos é que a forma dessa distribuição é influenciada pelos valores dos inputs já empregados e dos outputs já gerados em cada DMU. Os modelos foram construídos de acordo com o perfil geométrico da fronteira CCR, podendo ter forma côncava ou convexa conforme a natureza da variável que se quer distribuir. Mostramos que, sob essa hipótese de um formato geométrico específico para a fronteira, torna-se simples a distribuição do novo input (ou output) entre as várias DMUs, pois cada uma delas receberá a fração do novo input (ou output) que a torne eficiente, forçando sua posição sobre a fronteira pré-determinada.Palavras-chave: análise envoltória de dados; soma constante de variáveis; fronteira eficiente. AbstractIn this work we present two DEA (Data Envelopment Analysis) models in which a new input (or output) is to be fairly assigned to all units (DMUs -Decision Making Units) of a corporation. An interesting characteristic of those proposed models is that the way this distribution is influenced by both inputs and outputs used in each DMU. The models were built in accordance with the geometric shape of CCR frontier and they can be convex or concave, depending on the nature of the variables we intend to share. We show that, under this assumption, it becomes relatively easy the distribution of a new input (or output) among several DMU's, considering that each DMU will receive the new input (or output) fraction in order to become efficient, forcing its position on the specific frontier.

ABSTRACT. In this paper we present a brief overview of 10 years of research on Parametric Data Envelopment Analysis (DEA). We begin with a brief introduction to DEA. Then, we present the central problem of how to distribute a new and total fixed variable (input or output) across all DMUs (Decision Making Units) present in the analysis. We then tell the history of the developments in Parametric DEA, beginning with its first conception in 2002, moving onward to the current research. We conclude with a brief discussion of future perspectives.

This article develops a model for determining the optimal inspection schedule for a system which deteriorates according to a semi‐Markov process that progresses through three states: good, defective, and bad. A binary test is used, and false positives may occur. A true positive results in an action that reduces the likelihood of entering the bad state, but at most one such corrective action can occur during the lifetime of the system. Costs are associated with each inspection, each false positive, the corrective action, and the entrance into the bad state. Dynamic programming is used to compute the minimum expected cost, which is a function of the age of the system. The optimal inspection schedule is readily derived from this value function. Computational examples are provided. This model is appropriate for medical screening or for a mission where there is only one spare part.

In this work we investigate the relative efficiency of 34 Brazilian Landline Telephone Service companies using Data Envelopment Analysis with weight constraints in the input and output variables. We formulate two different models that take into account the performance of the companies with respect to the criteria defined by Brazilian National Agency of Telecommunications (ANATEL). We also illustrate the potential of efficiency improvement through the simulation of corporate Merger.