Rumour Detection and Analysis on Twitter

In recent years people have become increasingly reliant on social media to read news and get information, and some social media users post unsubstantiated information to gain attention. Such information is known as rumours. Nowadays, rumour detection is receiving a growing amount of attention because of the pandemic of the New Coronavirus, which has led to a large number of rumours being spread. In this paper, a Natural Language Processing (NLP) system is built to predict rumours. The best model is applied to the COVID-19 tweets to conduct exploratory data analysis. The contribution of this study is twofold: (1) to compare rumours and facts using state-of-the-art natural language processing models in two dimensions: language structure and propagation route. (2) An analysis of how rumours differ from facts in terms of their lexical use and the emotions they imply. This study shows that linguistic structure is a better feature to distinguish rumours from facts compared to the propagation path. In addition, rumour tweets contain more vocabulary related to politics and negative emotions.

Stop Words for Processing Software Engineering Documents: Do they Matter?

Stop words, which are considered non-predictive, are often eliminated in natural language processing tasks. However, the definition of uninformative vocabulary is vague, so most algorithms use general knowledge-based stop lists to remove stop words. There is an ongoing debate among academics about the usefulness of stop word elimination, especially in domain-specific settings. In this work, we investigate the usefulness of stop word removal in a software engineering context. To do this, we replicate and experiment with three software engineering research tools from related work. Additionally, we construct a corpus of software engineering domain-related text from 10,000 Stack Overflow questions and identify 200 domain-specific stop words using traditional information-theoretic methods. Our results show that the use of domain-specific stop words significantly improved the performance of research tools compared to the use of a general stop list and that 17 out of 19 evaluation measures showed better performance.