Overview

This section addresses several subjects.

• How time series are used in business and finance.

• What systems (FAME, S-Plus, SAS, KSQL) that are specialized for array processing do to support them.

• A brief discussion of what the relational vendors are offering, focusing on Oracle 8i. (The reason we are brief is that we view these offerings as evolving.)

• An annotated bibliography in time series for computer scientists.

Since much of this work was motivated by Shasha's consulting on Wall Street, we start with a trading example to serve as motivation. An analytical group discovers the desirability of "pairs trading." The goal is to identify pairs (or, in general, groups) of stocks whose prices track one another after factoring in dividends.

You can make money (lots has been made) because, for example, you may know that the two banks Chase and Citibank track each other (their difference is a stationary process). Now if, on a certain day, Chase goes up but Citibank doesn't, then buy Citibank and sell Chase. This is simplified: the model often involves a whole collection of stocks, and externalities such as dividends and business news must be taken into account. Typical challenge queries from such an application include the following:

• Correlate the price histories of two stocks or in general among many stocks and options. (For most traders, returns are more interesting than prices because they have better statistics: a stock that trends up over the years has an unstationary mean, but perhaps a stationary return.