兩部分計數迴歸模型的穩健與雙穩健M估計在大量零且缺失資料的應用

In the area of statistics, estimators are generally the products of interest to be developed and used for inference. Hence, those estimators must have nice qualities in order to be used for specific purposes and specific conditions. Several new estimators with nice properties will be developed, accordingly.The proposed semiparametric M-estimators will be able to produce better results in situations where they are needed. That is because some are double robust and others are more robust compared with the existing ones pointed in Lukusa et al. (2017). They can enable us to obtain results that are more efficient and inclusive compared with the existing ones.Moreover, the proposed methods can improved previous findings for instance Lukusa and Phoa (2019) Finally, this study can become a reliable reference for future research in health economic since there is no study related to missing data as far as we know.