Faculty News
Professor Vasant Dhar discusses why big data failed to predict the presidential election
—
Excerpt from Mashable -- "When you're predicting frequent versus infrequent events, it's a whole different ball game, when it comes to making predictions. So it's one thing if you have lots and lots of examples of someone having bought something or having done something and you're making a prediction, versus a situation where it's a one-time event that you're predicting, and so you don't have actual occurrences of winners and losers in presidential elections. There are so few."
Listen here
Listen here