Tools
Tools
Ali, Rao Waqar (2025) An NLP Powered Pipeline for Automated Construction and Sentiment Labelling of a Multi-Source Irish News. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (1MB) | Preview |
Abstract
The sheer increase in the digital news content poses tremendous challenges to the automated processing and analysis. The paper constructs and evaluates an end-to-end Natural Language Processing pipeline that targets the analysis of the sentiment of Irish news media. The most impactful element that occurred to me when I first ventured into the exploration of this research was the overwhelming amount of digital news that is being generated every single day by the media houses in Ireland. The project is a cumulative result of my attempt to create an advanced Natural Lan gauge Processing framework with an exceptionally narrow scope to send sentiments analysis to an Irish journalism environment. My constructed system incorporates five somewhat different methodological approaches, namely VADER to analyze the lexicon of a text, BERT to serve as a benchmark in text-to-text transformations, BER-Tweet to optimize on social media, and an ensemble model to use majority vote strategies [31, 30, 29, 28]. By automating the process of collecting data using the major Irish sources such as RTE, The Journal and The Irish Times using RSS feeds and highly sophisticated preprocessing methods and running the former model parallel I was able to form a very strong evaluation framework. What came out in my analysis was very eye opening. BERT was shown to be extremely consistent, which is used as an excellent standard of reference. Ensemble method attained an accuracy of 87.69 percent and Vader scored averagely at 63.08 percent. The same was demonstrated to be limited by both BER- Tweet and T5 with 60.00 percent and 49.23 percent accuracy, respectively. These results would offer objective, evidence based recommendations on the choice of the model as well as define precious performance standards in the case of the Irish news sentiment analysis.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Simiscuka, Anderson UNSPECIFIED |
| Subjects: | Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing H Social Sciences > HM Sociology > Information Science > Communication > Mass media |
| Divisions: | School of Computing > Master of Science in Artificial Intelligence for Business |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 02 Sep 2026 11:35 |
| Last Modified: | 02 Sep 2026 11:35 |
| URI: | https://norma.ncirl.ie/id/eprint/9774 |
Actions (login required)
 |
View Item |