In this paper, we investigate right-truncated count data models incorporating cavariates into the parameters. A regression method is proposed to model right-truncated count data exibiting high heterogeneity. The study encompasses the formulation of the proposed model, parameter estimation using an Expectation-Maximisation (EM) algorithm, and the properties of these estimators. We also discuss model selection procedures for the proposed method. Furthermore, a Monte Carlo simulation study is presented to assess the performance of the proposed method and the model selection process. Results express accuracy under regularity conditions of the model. The method is used to analyze the determinants of the degree of adherence to preventive measures during teh COVID-19 pandemic. in northern Benin. The results show that a right-truncated Poisson mixture model is adequate to analyze these data. Using this model, we conclude that age, education level, and household size determine an individual's degree of adherence to preventive measures during COVID-19 in this region.

翻译：本文研究了将协变量纳入参数的右截断计数数据模型。针对呈现高度异质性的右截断计数数据，我们提出了一种回归建模方法。研究内容包括：所提模型的构建、使用期望最大化（EM）算法进行参数估计，以及这些估计量的性质。我们还讨论了该方法的模型选择流程。此外，通过蒙特卡洛模拟研究评估了所提方法及模型选择过程的性能。结果表明在模型的正则性条件下估计具有准确性。该方法被用于分析贝宁北部地区COVID-19大流行期间预防措施遵守程度的影响因素。结果显示右截断泊松混合模型能够有效分析此类数据。基于该模型，我们得出结论：在该地区，年龄、教育水平和家庭规模是决定个体在COVID-19期间预防措施遵守程度的关键因素。