| Popis: |
Close price pattern is one of the widely applied technical indicators in market operations and trading. However, datasets or programming packages including calculation or labeling of close price patterns barely exist in public. With understanding of difficulties and flexibility of identifying a typical close price pattern, we developed a pattern recognition algorithm specially for close price patterns using Featured-based Dynamic Time Warping and Pattern Rule. The algorithm exhibits identified sequences with close resemblance of target financial patterns while incorporating some variations tolerable in human sense. |
