- Thesis:
- Bachelor in Mathematics 50%
- Author:
- Arian Gjini
- Title:
- Diskrete Variablen und partiell beobachtete Modelle
- Supervisor:
- Jan JOHANNES
- Abstract:
- In this bachelor’s thesis, an intensive mathematical investigation of three central econometric models is carried out: the Dichotomous Model, the Multiple Choice Model, and the Sample Selection Model. Through detailed analysis of these models, including their mathematical foundations and empirical applications, I open new insights into the modeling and analysis of categorical data and data influenced by sample selection. The work begins with the Dichotomous Model, a key tool for analyzing binary decision processes, and particularly highlights the implementation and significance of the Maximum Likelihood Estimation. This method is practically applied through the creation of a boxplot that visualizes the estimation of the parameter lambda, thereby illustrating the theory and emphasizing the importance of precise data visualization. The Multiple Choice Model is discussed as an extension of the binary model, enabling decision processes with multiple outcomes. This analysis deepens the understanding of model complexity and the flexibility of the logistic model for econometric investigations. The Sample Selection Model, the third main topic, addresses selection biases and thus expands the econometric method spectrum. The careful examination of this model underscores the necessity of accurate model specification and estimation. In summary, the work offers a deep insight into the mathematical modeling and analysis of important econometric models, emphasizes the importance of mathematical precision in research, and demonstrates the practical relevance of these models for analyzing real data.
- References:
- J.-P. Florens, V. Marimoutou und A.P. Feisolle. Econometric modeling and inference, Cambridge University Press, 2007.
| Statistics Group | Institute for Mathematics | Faculty of Mathematics and Computer Science | University Heidelberg |