Lexicon and Machine Learning Based Approach Sentiment Analysis of English News

Sentiment analysis, a subfield of natural language processing, has garnered significant attention due to its wide-ranging applications in understanding public opinion, market trends, and social media sentiment. The study investigates the effectiveness of various approaches, including lexicon-based methods and machine learning algorithm such as Multinominal Naïve Bayes. The research methodology involves both qualitative and quantitative approaches. They include data collection from an English news source, preprocessing techniques to clean and prepare the text data, and the implementation of machine learning models. Evaluation metrics including accuracy is employed to assess the performance of the models. It also includes sentence-by-sentence article texts analysis. The findings of this study contribute to the understanding of how different lexicon resources and machine learning algorithms can be combined effectively for sentiment analysis tasks. Furthermore, the research provides insights into the challenges and opportunities in sentiment analysis of news articles. Concluding, this thesis aims to advance the field of sentiment analysis by proposing a comprehensive framework that leverages both lexicon-based approaches and machine learning techniques for analysing the sentiment expressed in English news articles, thereby facilitating deeper insights into public opinion and discourse analysis.