Abstract

SALAZAR, JUAN CARLOS  and  BAENA ZAPATA, ARMANDO. ANALYSIS AND DESIGN OF EXPERIMENTS APPLIED TO SIMULATION STUDIES. Dyna rev.fac.nac.minas [online]. 2009, vol.76, n.159, pp.249-257. ISSN 0012-7353.

Simulation studies, when used in several areas of investigation, are quite useful to study the behavior of some phenomena in which different virtual situations are generated by the researcher using some specialized software. Robustness studies are rather common in statistic research; many of them are used to observe the behavior of an estimator under several hypothetical situations that could happen in practice. Due to the similarity among the studies of simulation and the experimental studies, the aim of this work is to propose the use of both the methodology and the design and analysis of experiments in the studies of simulation. As an example it is presented a simulation study called “Robustness of a three-state Markov model under different distributional specifications of the transition times” that was completed under the approach of the experimental design and using the analysis of variance and multiple regression model in order to prove the effect of the sample size, maximum number of visits and the type of distribution of the transition times on the estimation of the parameters in a three-state Markov model.

Keywords : Simulation studies; Design of Experiments; Generalized Linear Models; Stochastic Processes.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )