Mining Trading Rules from Stock Time Series Based on Charting Patterns
Jian Wu · Systems Engineering · 2004
This paper proposes a method by which investors can acquire some trading rules from stock time series based on their experience and intuition. It can recognize similar chart from history price series via pre-defined chart pattern and similarity preference. Partitioning similarity value into several intervals, and computing the return of each buy sell interval pair, some trading rules are obtained based on statistical hypothesis tests. This system can also be used to verify the validity of some technical chart patterns.