Data mining and AI techniques applied to forex trading domain

Abstract

The possibility of the existence of temporal relationships between a number of financial instruments and otherwise, such as the prices of FOREX Currency Pairs and/or commodities [1 ], and/or mentions and frequency of keywords and/or sentiment in popular sources of information [2] was explored. The discovery of these temporal relationships might greatly aide FOREX traders in trading with a net profit, by designing appropriate trading strategies which take advantage of such relationships and possibly also contemporary techniques such as technical analysis, market fundamental analysis and market sentiment analysis [3]. Data was acquired, cleaned and harmonised in order to be used in this project. Cross correlation techniques to discover temporal relationships were investigated and documented. Methods to exploit such relationships were also documented.