Computer scientists as the Massachusetts Institute of Technology have developed a machine learning algorithm that can predict the value of bitcoin. Principal investigator Devavrat Shah and his lab have been able to almost double their bit coin investment using their prediction model. They developed an algorithm which is capable of identifying patterns in vast amount of data to predict the trends of bitcoin value. After analyzing bitcoin price variations for five months, resulting in over 200 million data points, the researchers taught an algorithm how to identify the trends in bitcoin prices fluctuation. The algorithm is capable of automatically trading bitcoin by using two seconds worth of information to predict the value of bitcoin over the next ten seconds.
“Can we explain the price variation in terms of factors related to the human world? We have not spent a lot of time doing that. But I can show you it works. Give me your money and I’d be happy to invest it for you,” says Shah
While this algorithm has proven effective in capturing financial returns and reducing risk in the bitcoin market, it is not as helpful in finding the reasons behind the fluctuations. The algorithm is also limited in the time scale it can work over. With just a ten seconds glimpse of the future, the algorithm has some major restrictions. Shah and his team are hoping to expand the breadth of data collection in an effort to make the algorithm more effective.
Read More – MIT Computer Scientists Can Predict Bitcoin Price (MIT News, Adam Connor-Simons)