I am interested to see if I can validate the positive sentiment in social media for a particular company by any financial measure (Stock!). As far as I know, stock prices are quite sensitive by many factors such (expanding markets which in Apple Co. Case the market expansion in china was the main reason for its stock price jump).
So the question: Is it possible to reduce other major factors that affect stock prices and explore the relationship of the remaining fluctuations with social media sentiment?
The following papers deal with ontology driven data mining for stock price predictions: http://ceur-ws.org/Vol-1088/ProceedingsUDMIJCAI2013.pdf#page=54 and http://edocs.fu-berlin.de/docs/servlets/MCRFileNodeServlet/FUDOCS_derivate_000000002257/TR-B-12-07.pdf?hosts=local. They differ from your question as they deal with trends, that is future movements, and use stock news instead of social media but should still give an insight to the correlation of news and stock movements. You will certainly find some correlation between stock prices and social network data if your data is relevant. Problems might arise in context recognition due to the number of financial instruments or derivatives. A positive sentiment in context of the stock itself is quite the opposite to positive sentiment in regard to a short option on the stock and often hard to distinguish.
The interesting question is, whether social network sentiments follow the stock price or the stock price follows the sentiment. And going further it would be extremely interesting to find groupings of some form (real connections or similar skills, interests) that differ in that regard. So one group would initiate a movement in the stock price (trendsetter) and the other following a recognized stock movement (trendfollower). In the market these groups are typically large investors as trendsetters and private investors as followers. Finding such groupings again within social networks would be fascinating.
hi Arash, the paper below may not be particular to your interest. But hope it may give u some idea on how social media particularly Wiki shapes information environment in fanatical market.
Xu, S. X., & Zhang, X. M. (2013). Impact of Wikipedia on market information environment: Evidence on management disclosure and investor reaction. Mis Quarterly, 37(4), 1043-1068.
Yes! In the short run.This fact has been validated by some academic journal articles on volatility of the Nigerian stock exchange prices in the boom years of 2005-2009.In the boom period corporate dividend payments presented in the social media engenders financial signals that influence investors' buying and selling behaviour-
I believe that social media "plague" investors and their investment decisions. Certainly, social networks have accelerated the spread of information and at the same time affect the reaction of financial markets and their information efficiency. See if this paper may be useful to you.
http://www.hicss.hawaii.edu/hicss_47/bestpapers/Trading%20on%20Twitter_c_%20The%20Financial%20Information%20Content%20of%20Emotion%20in%20Social%20Media.pdf
Social network for stock information must be done with extreme caution, it is better to follow stock market or individual company information for effective stock market analysis
The following papers deal with ontology driven data mining for stock price predictions: http://ceur-ws.org/Vol-1088/ProceedingsUDMIJCAI2013.pdf#page=54 and http://edocs.fu-berlin.de/docs/servlets/MCRFileNodeServlet/FUDOCS_derivate_000000002257/TR-B-12-07.pdf?hosts=local. They differ from your question as they deal with trends, that is future movements, and use stock news instead of social media but should still give an insight to the correlation of news and stock movements. You will certainly find some correlation between stock prices and social network data if your data is relevant. Problems might arise in context recognition due to the number of financial instruments or derivatives. A positive sentiment in context of the stock itself is quite the opposite to positive sentiment in regard to a short option on the stock and often hard to distinguish.
The interesting question is, whether social network sentiments follow the stock price or the stock price follows the sentiment. And going further it would be extremely interesting to find groupings of some form (real connections or similar skills, interests) that differ in that regard. So one group would initiate a movement in the stock price (trendsetter) and the other following a recognized stock movement (trendfollower). In the market these groups are typically large investors as trendsetters and private investors as followers. Finding such groupings again within social networks would be fascinating.
An analysis of almost 10 million tweets from 2008 shows how they can be used to predict stock market movements up to 6 days in advance http://arxiv.org/abs/1010.3003
http://arxiv.org/abs/1010.3003
Seems that can be possible, I think justifications for the tweets ( to analysis) will however be of an issue. Tweets from the company, tweets from the public or interaction between both, what time are you going to capture the tweets, just before a product lunch? Hope this help.
No matter what, the result of such endeavor is going to be in qualitative term because it is a perception, not a documented data. Such endeavor however should be done with a lot of caution to avoid a misled and mis- information.
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Social Media