Quantifying implicit political intentions in parliamentary discourse - an integrated approach using semi-supervised learning and vector space information retrieval

Zejnilovic, LeidSturm, Niclas Frederic2022-06-172022-06-172022-01-202021-12-15http://hdl.handle.net/10362/140155Recent advances in Natural Language Processing and Information Retrieval have opened a new world of possibilities for the analysis of text. This study seeks to explore the possibilities of applying these techniques on political text, with a focus on quantifying intentions in parliamentary speeches and activities in Portugal. Combining vector space models and semi-supervised learning, a semantic search engine is able to extract meaningful metrics from text that help to identify political trends and quantify alignment with political issues.engNatural language processingPolitical scienceSemi-supervised learningInformational retrievalDiscourse analysisBusiness analysisQuantifying implicit political intentions in parliamentary discourse - an integrated approach using semi-supervised learning and vector space information retrievalmaster thesis202997375