The algorithm did pretty well using stock market data alone, predicting the shape of the stock market with 73.3 percent accuracy. But it did even better when the emotional information was added, reaching up to 86.7 percent accuracy.
----------------------------------------
Doesn't everyone try to correlate everything with the Dow?
Accuracy improves if we add just a teensy weensy bit of something else in this case adjectives aka emotions.
Using twitter is just one more form of a thousand other correlations.
They mention "emotions' using 72 adjectives. Instead of words in this case "adjectives" and instead we assign a number or better yet tweek the adjective with a numerical value.
âfriendly,â = 1
âpeeved,â = 2
âactive,â =3
âon edgeâ = 4
âpanicky,â = 5
so on and so forth.
At this moment in time, suppose the system works. What do we know for sure? It works until is doesn't.
Just sayin..
Probably since the article went to press, we have researchers correlating Twitter with gold, oil, etc.
suppose the entire thing is a PR stunt. NY has the highest % of active twitter users.
(In other news the vampire squid was behind the whole thing lmao)
