Not logged in.
Quick Search - Contribution
Contribution Details
Type | Master's Thesis |
Scope | Discipline-based scholarship |
Title | Forecasting Financial Time Series Based on Sentiment Analysis |
Organization Unit | |
Authors |
|
Supervisors |
|
Language |
|
Institution | University of Zurich |
Faculty | Faculty of Business, Economics and Informatics |
Number of Pages | 46 |
Date | 2016 |
Abstract Text | Previous research uses positive and negative word-lists to quantitatively measure the general tone of financial articles and study the effect of these sentiment variables on financial time series. This paper introduces a novelty feature construction method of financial documents based on natural language processing techniques and appraisal theory from a psychological point of view. I find that uncertainty emotion expressed in news articles can reflect the stock trading volume of individual firms. Besides, sentiment features like surprise and uncertainty have significant influence on cumulative abnormal returns during a period after the earning announcement day. These features give valuable information of the post-earnings-announcement drift by associate the drift direction with emotions expressed in financial articles. Keywords: Financial Textual Analysis, Post-Earnings-Announcement Drift, Appraisal Theory, Natural Language Processing, Sentiment Analysis. |
Export | BibTeX |