Predicting individual shares and stock market movements is one of the oldest professions and there are people far better qualified than me to assess the validity of the models used. Automating this process is also not new.
Tech-savvy traders have been sourcing data from reports, press releases and corporate Web sites for years. Genuine advocates of automated predictive models use powerful computers to speed-read content and then letting the machines decide what it all means for the markets.
However, using Twitter is still a relatively new phenomena. In this and my next blog, I will explore how two studies I have come across can inform our thinking on social media and share price.
Twitter mood predicts the stock market. (Johan Bollen, Huina Mao, Xiao-Jun Zeng)
The Bollen et al study results indicated that the accuracy of DJIA predictions from Twitter can be significantly improved by the inclusion of specific public mood dimensions. They claimed an accuracy of 87.6% in predicting the daily up and down changes in the closing values of the DJIA and a reduction of the Mean Average Percentage Error by more than 6%.
The study investigated whether measurements of collective mood states derived from large-scale Twitter feeds are correlated to the value of the Dow Jones Industrial Average (DJIA) over time. They analysed text content of daily Twitter feeds with tracking tools that measure positive vs. negative mood and mood in terms of 6 dimensions (Calm, Alert, Sure, Vital, Kind, and Happy).
Now I’m not convinced by this study.
It seems to me, if the mood is being reflected in the Twittersphere, everyone probably already knows how it will move markets, so there isn’t a whole lot of competitive advantage. Unless of course, it’s insider information, in which case and putting aside the regulatory and moral minefield that comes with this hypotheses; Investors would need a high risk tolerance to act on it without verification and the opportunity cost of seeking verification would render the Twitter news obsolete.
If that isn’t enough to be wary of, as a researcher posted on Quora “There's also a paradox: once there are enough tweets about the inside information it's no longer really inside information. Twitter is probably best used for assessing the herd mentality. One strategy might be to look at all the dumb stuff people are saying and then bet the opposite.”
In my next blog, I will look at the sales effect of the volume of positive, negative, and neutral online communications captured by social media monitoring technology and classified by automated sentiment analysis.
Tech-savvy traders have been sourcing data from reports, press releases and corporate Web sites for years. Genuine advocates of automated predictive models use powerful computers to speed-read content and then letting the machines decide what it all means for the markets.
However, using Twitter is still a relatively new phenomena. In this and my next blog, I will explore how two studies I have come across can inform our thinking on social media and share price.
Twitter mood predicts the stock market. (Johan Bollen, Huina Mao, Xiao-Jun Zeng)
The Bollen et al study results indicated that the accuracy of DJIA predictions from Twitter can be significantly improved by the inclusion of specific public mood dimensions. They claimed an accuracy of 87.6% in predicting the daily up and down changes in the closing values of the DJIA and a reduction of the Mean Average Percentage Error by more than 6%.
The study investigated whether measurements of collective mood states derived from large-scale Twitter feeds are correlated to the value of the Dow Jones Industrial Average (DJIA) over time. They analysed text content of daily Twitter feeds with tracking tools that measure positive vs. negative mood and mood in terms of 6 dimensions (Calm, Alert, Sure, Vital, Kind, and Happy).
Now I’m not convinced by this study.
It seems to me, if the mood is being reflected in the Twittersphere, everyone probably already knows how it will move markets, so there isn’t a whole lot of competitive advantage. Unless of course, it’s insider information, in which case and putting aside the regulatory and moral minefield that comes with this hypotheses; Investors would need a high risk tolerance to act on it without verification and the opportunity cost of seeking verification would render the Twitter news obsolete.
If that isn’t enough to be wary of, as a researcher posted on Quora “There's also a paradox: once there are enough tweets about the inside information it's no longer really inside information. Twitter is probably best used for assessing the herd mentality. One strategy might be to look at all the dumb stuff people are saying and then bet the opposite.”
In my next blog, I will look at the sales effect of the volume of positive, negative, and neutral online communications captured by social media monitoring technology and classified by automated sentiment analysis.