Biomedical subjects

M Elketroussi

Publications and source records attributed to M Elketroussi.

4 recordsLinked to original sources

PubMed1994

Optimization of simulation models with GADELO: a multi-population genetic algorithm.

In this paper, a new Genetic Algorithm based on the Dynamic Exploration of Local Optima (GADELO) was used to estimate the parameters of the MRD (Micro-population model of Risk-group Dynamics) micro-population model for smoking cessation by minimizing a deviation function between the model's predictions and the smoking cessation data of the Multiple Risk Factor Intervention Trial (MRFIT). The efficiency and accuracy of the GADELO estimations were consistently superior to those obtained using the standard genetic algorithm or the simplex algorithm of Nelder-Mead.

Algorithms↗

PubMed1992

Time trends of smoking cessation: a micro-population computer simulation model.

The Micro-population model of Risk-group Dynamics (MRD) approaches smoking behavior at the level of the individual and integrates physiological and social factors to describe the evolution of behavior change in the population. MRD is innovative in several ways: (1) the model describes mathematically the interactions among these behavioral factors; (2) the model accounts for both the variability of these factors among different persons and the universality of basic rules describing these factors in all individuals; and (3) the model can be applied to various types of populations and a wide range of intervention strategies. MRD combines the physiological, psychological and social determinants into a hazard function for relapse to smoking. This hazard function is then organized into a three term expression incorporating: a baseline hazard characteristic to each individual, a decreasing term for the diminishing aspect of the initial hazard and an effect of external interventions. The model gives promising results when applied to the Multiple Risk Factors Intervention Trial (MRFIT) data using the assumptions of a Weibull distribution for the baseline hazard, a negative exponential for the decrease in the initial hazard and a constant intensity for the external intervention.

Algorithms↗

PubMed1989

Mathematical model for addiction: application to multiple risk factor intervention trial data for smoking.

Describes habituation and addiction, both psychological and physiological, using the simple equations of the mathematical model of ideodynamics. The parameters in these equations were optimized to smoking data from the Multiple Risk Factor Intervention Trial (MRFIT) program. With only 4 constant parameters, it was possible to calculate accurate time trends for recidivism to smoking among quitters, time trends for secondary cessation among recidivists, and final percentage of smokers in a population with both recidivism and secondary cessation occurring simultaneously. These same parameters further permit predictions for the long-range success of intervention programs to decrease substance dependency. Ideodynamics can also predict time trends of public opinion based on stories in the mass media.

Behavior Therapy↗

PubMedDate not supplied

Time trends of smoking cessation analyzed with six mathematical survival models.

In this paper, six mathematical models were applied to model time trends of smoking cessation. Both statistical and non-statistical methods were used and included the exponential, ideodynamic, log-logistic, Pareto, sickle and Weibull models. All models included the possibilities of both permanent abstinence and relapse to smoking. Time trends from all models were compared with data from the Multiple Risk Factor Intervention Trial (MRFIT) program. The Pareto, log-logistic, Weibull and ideodynamic models yielded satisfactory fits to the data while the sickle and exponential models did not. Even though the data used in this paper were not sufficient to distinguish among these four models, the methodology will be useful for further narrowing the model choices as additional data for the testing become available.

Adult↗