In the simplest case, we are given a collection of sequences fS1; S2; : : :; Sng, and a query sequence Q, each
of the same length. Our problem is to nd that sequence Si whose distance from Q is the minimum, where
distance" is de ned by the energy" of the di erence of the sequences; i.e., D(S; T) = R 1
0 S(t) T(t)2
dt.
For instance, the Si′s might be records of the prices of various stocks, and Q is the price of IBM stock,
delayed by one day. If we found some Si that was very similar to Q, we could use the price of the stock Si
to predict the price of IBM stock the next day, Notes:
Do not try this at home. Anything easy to mine about stock prices is already being done, and the
market has adjusted to whatever knowledge can be gleaned.
Sequence matching is a great opportunity to violate the Bonferroni principal, since there has to be a
closest sequence." For instance, a famous mistake was looking in the UN book of world statistics to 数据挖掘研究院
nd the statistic that best predicted the Dow-Jones average. It was cotton production in Bangladesh."
资料全文下载
