Aspect based Sentiment Analysis for Urdu

Aspect based sentiment analysis (ABSA) is an important task that aims to determine the polarities of the given aspects in its context. Most of the study on ABSA mainly focuses on resource-rich languages, such as English and Chinese, there is little relevant research for the Urdu language. In this paper, we establish a fine-grained aspect-based sentiment analysis dataset for Urdu. This dataset contains Urdu reviews of restaurants and laptops with specific aspect words and polarities. Further, we conduct different experiments on the dataset to build an ABSA system using state-of-the-art algorithms. Our work aims to build reliable baselines for ABSA in Urdu.