Patterns in financial markets: Dynamic time warping

Abstract

This work project introduces the performance of the algorithm Dynamic Time Warping amidst trading strategies in the financial markets. The employed procedure allows the comparison between any two sequences of data with different time lengths. Different features for the method were implemented, although those did not improve its promptness or accuracy in the outcomes obtained. Two potential investment strategies are presented within this theme. One yielded satisfactory outcomes whilst the other resulted in inconsistent values. The results point to the possible existence of patterns in the Equity Indexes’ behaviour, as well as their distortion across the time axis.