Main Content
The value and relevance of these types of sentiment indices is unexplored territory and questions about their value and accuracy are largely unanswered.
Data mining and stock market information flows
Data mining and stock market information flows
This study explores information mining methods that can be used to produce market sentiment information and will assess the impact of existing methods on the behavior of market prices.
The major metrics will be based on concepts related to the Shannon entropy theory and cross entropy. The project uses various data sets including Sentiment Data from The Securities Industry Research Centre of Asia-Pacific (SIRCA). This data set is sourced from Thomson Reuters who employ text analysis data mining techniques to screen market news feeds to assess whether company information is positive or negative and combines them to produce sentiment indices.
Research questions
- Do they have information value?
- How do they link to stock price movements and volatility?
- Do they lead or lag stock returns?
- Do positive movements in the sentiment index lower volatility?
Researchers
- Professor David Allen, (Chief Investigator).
- Dr Abhay Kumar Singh, (Post-Doctoral Fellow).
- Professor Michael McAleer, The Tinbergen Institute, Erasmus University, (Partner Investigator).