Persian Causality Corpus (PerCause) and the Causality Detection Benchmark

Recognizing causal elements and causal relations in the text is among the challenging issues in natural language processing (NLP), specifically in low-resource languages such as Persian. In this research, we prepare a causality human-annotated corpus for the Persian language. This corpus consists of 4446 sentences and 5128 causal relations. Three labels of Cause, Effect, and Causal mark are specified to each relation, if possible. We used this corpus to train a system for detecting causal elements’ boundaries.Also, we present a causality detection benchmark for three machine-learning methods and two deep learning systems based on this corpus. Performance evaluations indicate that our best total result is obtained through the CRF classifier, which provides an F-measure of 0.76. In addition, the best accuracy (91.4%) is obtained through the BiLSTM-CRF deep learning method