Fuzzy modeling of stock trading with fuzzy candlesticks. (1st March 2018)

• Record Type: Journal Article
• Authors: Naranjo, Rodrigo; Arroyo, Javier; Santos, Matilde
• Is Part Of: Expert systems with applications. Volume 93(2018)
• Page Start: 15
• Page End: 27
• Publication Date: 2018-03-01
• Subjects: Trading -- Fuzzy logic -- Candlesticks -- Forecasting -- Pattern recognition -- Stock market; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2017.10.002
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A methodology to model trading rules for candlestick patterns using fuzzy logic. The fuzzy trading system adapts three well-known candlestick patterns to fuzzy rules. The fuzzy system performance is tested in two portfolios from different stock markets. The fuzzy system is more profitable than a crisp version of the candlestick rules. The fuzzy trading is more risk averse than the Buy and Hold (B&H) strategy used as benchmark. Abstract: This paper proposes a methodology to detect candlestick patterns in a stock trading system using fuzzy logic. The fuzzy approach make possible to account for the vagueness and uncertainty of the pattern features. Even more, the use of fuzzy rules allows it to include that uncertainty into a trading decision system that not only advises the investor on when but also on how much capital to invest. In this way, the intelligent system helps the experts to use their knowledge, i.e. the candlestick-based rules, in a more natural and realistic way than the standard one based on crisp rules. In the paper we have illustrated this methodology with the generation of a fuzzy trading system that uses three well-known candlestick patterns that have been fuzzified. The performance of this intelligent stock trading system is tested in two portfolios of different stock markets, Nasdaq-100 and Eurostoxx, and it is compared against its crisp counterpart and the classical Buy-and-Hold trading strategy. Our fuzzy candlestick-based trading system not only improves the pattern recognition with respect to its crisp version, but it also provides promising results since it exhibits a more stable behavior in the markets analyzed, and obtains more profits in a less risky way than the other trading systems considered.