On the Relationship between IPO News Sentiment and IPO Oversubscription

  • 黃 昱翔

Student thesis: Doctoral Thesis

Abstract

SUMMARY With the continuous development and progress of the Internet various programming models and algorithms have also rapidly emerged In recent years big data artificial intelligence machine learning deep learning etc are quite popular topics Using deep learning we can extract the data we need and turn these text data into quantified scores or data for research The research is to use a deep learning model to convert every piece of IPO news reported by the media into an IPO news sentiment polarity score (i e positive or negative sentiment) and to explore the relationship between IPO news sentiment and IPO oversubscription The study found that the more positive the IPO news sentiment the greater the chance of causing IPO oversubscription; the more negative the IPO news sentiment the lower the chance of causing IPO oversubscription Afterwards we multiplied by the score of each sentence with the modal (the excitation level) of each sentence Based on results we found that the positive association between IPO oversubscription and sentiments extracted from relevant news remains statistically positive
Date of Award2020
Original languageEnglish
SupervisorMeng-Feng Yen (Supervisor)

Cite this

'