Mining candlesticks patterns on stock series: A fuzzy logic approach

Abstract

Candlesticks is a technique used in financial time series in order to forecast future market performance. With candlesticks patterns, traders build active trading strategies in order to buy, sell or hold securities. The process is based on a preliminary stage which consists in identifying individual basic shapes on time series. Identifying candlesticks basic shapes is easy for a human, but recognizing complex patterns is hard because a lot of data is available. In this paper a data mining model for building active trading strategies (using candlesticks assumptions) is proposed looking for frequent itemsets on symbolic stocks series. Model validation is achieved with real data from New York Stock Exchange. © 2009 Springer.