ABSTRACT
The random process everyone would like to be able to predict is the fluctuation of the share price of a particular stock. Honest stock traders’ skills in predicting this behavior rely on knowledge of the market in general and a study of the behavior of the stock in the past. Dishonest traders use “inside information” to concoct superior models, and the public at large is righteously infuriated when they are exposed. The so-called Yule-Walker equations enable the computation of an autoregressive-moving-average (ARMA) random process with known autocorrelations. To predict the weather, the stock market, the World Series outcome, the reliability of a circuit, and the lifetime of a machine, we seek a rational way to formulate a guess about something that is uncertain. Our basic task is to learn how to predict the values of random processes.
