Twitter: half of the Prophet, half mentally (graphs)

Prev close 3 points, who guard who attack?

2 of Wall Street, marked success?

Stock market what tips, Twitter to give you a solution.

Research has shown that fluctuate on Twitter can be used to predict the market fluctuations.

Comparison with body Twitter volatility curves and the Dow Jones industrial index changes of three days, find 86.7% similarity.

Indiana University's metering sociologist JohanBollen cannot help but sigh, this is an amazing result! his papers quickly become the focus of online Chase.

Obviously, the friends were pursued, not paper, and money.

Bollen and his students wool Huina (transliteration) is the chance to discover this conclusion.

Early research shows, celebrity blog has a tendency to incite emotion. This is a celebrity, the halo effect, herding. You can view those crazy mob are opinion leaders led by the nose. However, Twitter's remarks about the movies that can predict the box office. Through the open source of emotional OpenFinder Tracker, you can very clearly emotional change of keyword analysis, and summarize the positive evaluation and negative evaluation.

However, did not meet Bollen, and creates a mood barometer, ready to publish data, generating curves.

He specification for psychology in the usual mood table to describe this process. Similar tools commonly used in pharmaceutical research and doping control.

Of course, the questionnaire is commonly used to measure the extent of the respondents emotion, typically corresponding to 72 species of adjectives.

For example, friendly, crazy, positive, painful, crash ... And from the six dimensions measuring emotions. The six dimensions are: calm, sensitive, kind, important, and happiness.

Bollen and his colleagues through the Google database, retrieving those are often used to 72 species of adjectives, and incorporated in their dictionary.

And then they collected 270 000 Twitter in February 2008 to December, a total of 980 million micro-Bo, from which the filtered content of the show emotions (such as "I think," I want "), for a set of test data.

Bollen will this process graphically compared to doctor.

In their eyes, Twitter is a psychopath, but they have to do is pass 6 dimensions measuring their emotional state.

In the spirit of good detection, analysts will pass the number of media events to grasp the public mood.

For example, the 2008 presidential election and Thanksgiving, results show that Twitter is rendered before the election, but anxiety neurosis election day is more calm, happy, peaceful. To 5 November results announced, as usual. Thanksgiving Day, Twitter is like a happy child, excited in Cadence.

So, for future events, Twitter will be why through the comparison of national sentiment and the Dow Jones industrial index, the researchers found that even if the stock market heats up, calm is the only emotion.

However, the forecast than the actual situation, 4 days in advance, 3.

The result is very Bollen.

However, there is still no on Twitter and predict the future relevance of accurate inferences. But the test found to Twitter modeling can simulate stock market fluctuations.

This data model is a good simulation of the stock market prediction accuracy of curves, 73.3%.

If you join the emotional information, accuracy should be as high as 86.7%. Bollen stressed that emotional information critical for prediction, this simplified model is not the best predictor of stock market, if you have a lie detector that will greatly improve the accuracy of the prediction. In fact, merely by virtue of Twitter emotional information forecasting 2008 interim stock market is very accurate. This is the most important forecast because of presidential elections, war on terror than 1930 more terrible financial crisis, if the prediction can touch the bottom line of the Dow Jones, things that are good in the future.

Bollen, Twitter reflects the investor's mind, but this is another analysis cannot be captured, because that is the emotional information, not the financial statements.

HP's lab analysts Sitaram Asur believes that the results are exciting, but still need to examine the relevance. He does not believe that the emotional volatility Twitter will guide the stock market. Because not every Twitter are shareholders. Correlation between size is a critical issue, if not careful, it is most likely due to the arbitrary and make a mockery of the research results. Bollen also agreed with this view, there is no denying that this prediction is uncertain, but the prospects for research, he is very optimistic. Through your mouth, determine what the money will become more and more rely on spectrum.