Loading...

Estimation and Inference of Heterogeneous Treatment Effects using Random Forests

Ghodrati, Mohsen | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57465 (02)
  4. University: Sharif University of Technology
  5. Department: Mathematical Sciences
  6. Advisor(s): Haji Mirsadeghi, Mir Omid
  7. Abstract:
  8. Sometimes, the relationship between two random variables is not a correlation, but, a causal relationship where only changes in one of the two variables can lead to changes in the second one. The importance of examining such cause-effect relationships lies in the fact that sometimes, beyond just knowing the correlation between two variables, we are looking for an active effect on one, and knowing whether or not other correlated variables will also change with the change of this variable is of key importance. Such a process of exploring and analyzing cause-effect relationships is called causal learning. Based on this, in this thesis, we will first review the methods of causal inference, as a crucial part of causal learning. Therefore, conditions on which machine learning methods can assist causal inference are studied and determined. Assuming these conditions, a couple of methods for estimating the Average Treatment Effect are introduced. We will then conclude the first half of this thesis by focusing on estimating methods with asymptotic inference possibilities. Meaning that we desire methods that not only are unbiased but also have competitive convergence rates. In the second half of this thesis, we will examine the problem of estimating the effect of treatment in a heterogeneous setting, that is, a setting where there is a variety of quantities and methods of treatment among the people of the same sample. Following this path, as we will see, parametric methods are not suitable even though they are fast. Because, in cases where the problem's settings do not match with such a parametric modeling, bias in the final estimation is unavoidable. Thus, we will be paying attention to non-parametric methods, such as random forests, and introduce two growing patterns suitable for causal inference which are both unbiased estimators and guarantee convergence to normal
  9. Keywords:
  10. Causal Inference ; Nonparametric Method ; Heterogeneous Treatment Effect ; Average Treatment Effect ; Generalized Random Forests

 Digital Object List

 Bookmark

...see more