Aspect Based Sentiment Analys For Urdu Language

Sentiment analysis is an important research direction in natural language processing.Aspect-based sentiment analysis is a fine-grained sentiment analysis,which aims to distinguish the sentiment polarity of each specific aspect/target in a given sentence.It is a hot research direction of text sentiment analysis.Existing research mainly focuses on resource-rich languages,such as English,Chinese,etc.,and there is less research on low-resource languages.Urdu is the national language of Pakistan.It is the first language of 70 million people,spoken by about 104 million people,and is distributed in Pakistan and India.Even Urdu is a widely used language.The lack of natural language processing resources and related data sets hinders sentiment analysis research in Urdu.Sentiment analysis research is relatively speaking,and mainly focuses on the sentiment judgment of document and sentence level.Urdu’s fine-grained sentiment analysis data set and related algorithm research are basically in a blank state.The main work of this paper includes:(1)This paper constructs an aspect-based sentiment analysis and annotation data set for Urdu.The data set contains Urdu reviews of restaurants and laptops.The restaurant data set contains 2,951 reviews and the laptop data set contains 4,721 reviews.For each comment,fine-grained sentiment labeling is carried out,including aspect and polarity labeling.(2)This article establishes a benchmark for the Urdu aspect-based sentiment analysis method.On the generated Urdu data set,this paper carried out alternative experiments,realized 11 fine-grained sentiment analysis methods,and compared the experimental results.These results can be used as the basis for future research in this area.