A computational implementation of stock charting: abrupt volume increase as signal for movement in New York Stock Exchange Composite Index

摘要

In this case study in knowledge engineering, data mining, and behavioral finance, we implement a variation of the bull flag stock charting heuristic using a template matching technique from pattern recognition to identify abrupt increases in volume in the New York Stock Exchange Composite Index. Such volume increases are found to signal subsequent increases in price under certain conditions during the period from 1981 to 1999, the Great Bull Market. A 120-trading-day history of price and volume is used to forecast price movement at horizons from 20 to 100 trading days.