Exploring causal directions through word occurrences: semi-supervised Bayesian classification framework

Determining causal directions in sentences plays a critical role in understanding the cause-and-effect relationship between entities. In this paper, we empirically show that word occurrences from several Internet domains resemble the characteristics of causal directions. Our research contributes to the knowledge of the underlying data generation process behind causal directions. We propose a two-phase method: (1) a Bayesian framework that generates synthetic data from posteriors by incorporating word occurrences from Internet domains, and (2) a pre-trained BERT model that utilizes the semantics of words based on context to perform classification. The proposed method demonstrates an improvement in performance for Cause-Effect relations in the SemEval-2010 dataset compared to random guessing.