Abstract:
The objective of this research is to compare the estimation methods of parameters in logistic regression model with outliers . The estimation methods are Weight Maximum Likelihood Method of Croux and Haesbroeck, Weight Maximum Likelihood Method of Rousseeuw and Christmann, Weight Maximum Likelihood Method of Daniel and Maximum Likelihood Method of Hobza, Pardo and Vajda of parameter. There are two level of outliers, mild and extreme, and four proportions of contamination 0.00, 0.05, 0.10 and 0.15. Mean Absolute Percentage Error (MAPE) is the criterion of comparison. The sample sizes are 20, 50,70,80,90 and 100. Coefficient are, and . Distribution of independent variable are normal and exponential. Simulation data and Monte Carlo method are used to compute MAPE. The experiment was repeated until the difference of parameters between of iteration and less than 0.0001 under each situations. Results of the research are as follows: In most case, Weight Maximum Likelihood Method of Daniel have smallest MAPE ,especially when sample sizes is 20 . Next Weight Maximum Likelihood Method of Rousseeuw and Christmann and Maximum Likelihood Method of Hobza, Pardo and Vajda and Weight Maximum Likelihood Method of Croux and Haesbroeck repectively. MAPE of multivariate normal of independent less than exponential. MAPE of parameters increase when level of outliers and proportion of outliers contamination increase but they decrease when sample sizes increase. The level of outliers, proportion of outliers contamination and sample sizes have effected on parameters estimates. But differentness of coefficient have not effected on parameters estimates.