• Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint
Share this Page URL
Help

Chapter 3. Patterns and Predictions > High-Tech Data Mining

High-Tech Data Mining

The Motley Fools are not the only ones torturing the data and looking for price patterns that supposedly predict future returns. It has been a trend, over the past couple of decades, for people with highly quantitative skills to model the stock market. Engineers, physicists, computer programmers, and others like to run the data through their models. (By the way, my early education and job experience was in electrical engineering.) As a finance faculty member, I periodically get a call from other faculty or doctoral students in these quantitative areas looking for stock data to run using Kalman filters, chaos theory, neural networks, or other “kitchen sink” models.

People with high-tech skills are called quants in the investment industry. These quants build sophisticated computer models that try to determine trends and associations between many variables in the market. Although their models and computer programs are elaborate, they usually succumb to the same problems as the less sophisticated data miners. That is, they end up with models that have no working basis. Therefore, they are usually just a quantification of the randomness in the sample of mined data. Since there is no reason to believe that future randomness matches the randomness of the past, there are no grounds to bet your hard-earned dollar on them.


PREVIEW

                                                                          

Not a subscriber?

Start A Free Trial


  
  • Creative Edge
  • Create BookmarkCreate Bookmark
  • Create Note or TagCreate Note or Tag
  • PrintPrint