Monthly Archives: April 2006

Predictive Models – Extreme Stories of Applications

Prediction always facinates me and big time prediction, why they did, how they did, and how it held up or failed big time is a facinating story of human interest, needs, application, and all the story behind the application or failure of it. Take a look at these applications.

The first one is very interesting and amazingly it predicts only 5% of the male behavior and yet good enough for NOVO, the elite program of science, culture, and living organizms, to bring out this as a very important contribution to science. No pick’n; Still, as a scientist, it makes you wonder what really goes into good predictions, important predictions and why it attracts great attention.

Who would believe that the differece between the index finger and the ringer can predict the man-like behavior vs. woman-like behavior (have to be careful about phraseology and its use; these are not mine); we will see what it means little later.

See http://www.ualberta.ca/~allie/Bailey_andH04.pdf, who show that there is discernible relationship between the ratio of index finger to the ring finger in relationship with aggressiveness of an individul; this is not useful to predict the women aggressivenes. According to the researchers, finger lengths explain about five per cent of the variation in the physical aggressivness of individuals’ personality measures (not emotional such as anger, cursing, and hostile behaviour), “…so research like this won’t allow you to draw conclusions about specific people. For example, you wouldn’t want to screen people for certain jobs based on their finger lengths,” Hurd, one of the author of the above article says. Also, they authors say that this predicts behaviour of men, not women. “But finger length can tell you a little bit about where personality comes from, and that’s what we are continuing to explore.”

Talking about PBS, a great interview with nobel laureates and financial market’s inner circle of people about “controling uncertainty and risk” and the huge attraction and possibilities of catastrophic failures when you use predictive equations, is here (it is about marketing stupid and the rest is about hindsight) : http://www.pbs.org/wgbh/nova/transcripts/2704stockmarket.html . This is a nice example of using mathematical and statistical algorithmic approach to predict the stock market. The trillion dollar formula is given here: http://www.pbs.org/wgbh/nova/stockmarket/formula.html. This trillion dollar formula that not only got Nobel price to the inventor, made lots of money to the investment company Salmon Brothers, and eventually failed big time, trying to take the whole world economy with it in 1998. Prediction is prediction and you can not bet on it that it will work 100% of the times and as the story goes eventually when crashing of property prices in Thailand and Bankruptcy claims of Russian government investor’s ran for liquidity crashing the business strategy, of investing in markets around the world and fully relying on Black-Scholes formula. The formula did not include how extreme events (catastrophic events) and financial behaviours that may be termed as “outliers” will shock the formula. It is a great reading to go through the http://www.pbs.org/wgbh/nova/transcripts/2704stockmarket.html.

–more to come

Blog contents copyright © 2006 Nethra Sambamoorthi;