Human-annotated dataset for social media sentiment analysis for Albanian language

Abstract

Social media was a heavily used platform by people in different countries to express their opinions about different crises, especially during the Covid-19 pandemics. This dataset is created through collecting people's comments in the news items on the official Facebook site of the National Institute of Public Health of Kosovo. The dataset contains a total of 10,132 comments that are human-annotated in the Albanian language as a low-resource language. The dataset was collected from March 12, 2020, and this coincides with the emergence of the first confirmed Covid-19 case in Kosovo until August 31, 2020, when the second wave started. Due to the scarcity of labeled data for low-resource languages, the dataset can be used by the research community in the field of machine learning, information retrieval, affective computing, as well as by the public agencies and decision makers.

Keywords

• Sentiment analysis
• Machine/deep learning
• Affective computing
• NLP
• Text classification