科学美国人60秒 SSS 2014-08-12(在线收听

 This is Scientific Americans 60-Second Science. I am Christopher Intagliata. Got a minute?

 
Wanna know what the most popular movies, albums and video games are? Ask the internet. But don't Google it. Check what other people Googled. That's the theory behind Google Flu Trends—which crunches search data to forecast flu prevalence. With admittedly mixed results.
 
Nevertheless, researchers have applied the idea to the stock market. They found that when searches related to business and politics go up, the market tends to take a dive. Queries for words like "election," "senate," and "CEO" are predictive of the downward trend—but not terms like "church," "earthquake" and "geometry."
 
Why the correlation? The researchers say investors and the public may be more likely to search for information on business and government when they're concerned about the economy. And if confidence in the economy drops, so might stock prices. The findings appear in the Proceedings of the National Academy of Sciences.
 
Over time, the correlation weakened—a sign that traders may already be using Internet search data in automated trading algorithms.
 
"Generally speaking, if an opportunity like this is known, then the effect will diminish, and that's what we think we're seeing in the data we looked at, from 2004 to 2012." Lead author Chester Curme, a research fellow at the Warwick Business School.
 
And before you head over to Google Trends, he says, keep this in mind: "If somebody knows that other traders use this method, then a predatory trader could manipulate search volumes on Google artificially in order to deceive other traders." As they say in the biz: past performance does not guarantee future results.
 
Thanks for the minute, for Scientific Americans 60-Second Science. I am Christopher Intagliata.
  原文地址:http://www.tingroom.com/lesson/sasss/2014/8/278168.html