Research
Working Papers
DoubleMLDeep: Estimation of Causal Effects with Multimodal Data
Sven Klaassen, Jan Teichert-Kluge, Philipp Bach, Victor Chernozhukov, Martin Spindler, Suhas Vijaykumar
Abstract: This paper explores the use of unstructured, multimodal data, namely text and images, in causal inference and treatment effect estimation. We propose a neural network architecture that is adapted to the double machine learning (DML) framework, specifically the partially linear model. An additional contribution of our paper is a new method to generate a semi-synthetic dataset which can be used to evaluate the performance of causal effect estimation in the presence of text and images as confounders. The proposed methods and architectures are evaluated on the semi-synthetic dataset and compared to standard approaches, highlighting the potential benefit of using text and images directly in causal studies. Our findings have implications for researchers and practitioners in economics, marketing, finance, medicine and data science in general who are interested in estimating causal quantities using non-traditional data.
Talks
Price Elasticity Estimation using Image and Text Data
Sven Klaassen, Jan Teichert-Kluge, Victor Chernozhukov, Martin Spindler
Abstract: This paper explores the use of text and image data in causal inference settings for estimating price elasticities using the double/debiased machine learning approach. The paper discusses the challenges of using text data for causal inference and the need for suitable methods to adjust for high-dimensional unstructured confounders. We present an empirical analysis of causal inference using text and image data and compare the results with standard approaches relying only on tabular data to estimate price elasticities. The study finds that our model can improve upon the limitations of the standard approach. The paper contributes to the literature on causal inference by demonstrating the potential of text and image data.
Other
Causal Machine Learning with Deep Learning approaches using the DoubleML Framework
Moritz Sundermann, Jan Teichert-Kluge
Abstract: In this paper, the authors present an object-oriented software implementation for Deep Learning approaches in the Double Machine Learning framework. The objective of this work is to verify the implementation and to examine the behaviour of the implemented methods in various data settings using a simulation study. In addition to theoretical explanations, a case study with observational data is conducted in order to demonstrate the practical applicability of the proposed methods in a real-world causal inference problem. Finally, all results and findings of this thesis are summarized and an outlook on directions for future research is given.
Not published.