On the Profitability of Scalping Strategies Based on Neural Networks
We analyze the potential of unsupervised neural networks when they are employed to support intraday trading activity on financial markets. Several time frequencies have been considered: from five minutes to daily trades. At the current stage our major findings may be summarized as follows: a) unsupervised neural networks are helpful to localize profitable intraday patterns, and they make possible to achieve higher performances than common trading rules; b) trading strategies based on neural networks make exploitable with profits almost continuous trades (i.e. scalping), until transaction costs maintain below proper thresholds.
KeywordsNeural Network Transaction Cost Trading Strategy Trading System Time Frequency
Unable to display preview. Download preview PDF.
- 4.Deboeck, G.J.: Modeling non-linear market dynamics for intra-day trading (1999), Available on-line at: http://www.dokus.com/PapersontheWeb/intradaymodel.htm
- 5.Graifer, V.: How to Scalp Any Market, Reality Trading (2005)Google Scholar