In this paper we consider a simple trading system (TS) based on a set of Technical Analysis (TA) indicators. Their peculiarity is the dependence on the time-window widths used to calculate them. To attempt to improve the performances of the TS, we optimize these parameters (that is the time-window widths) by the Particle Swarm Optimization (PSO), which is a metaheuristic used to solve global optimization problems. The use of PSO is necessary since the involved optimization problem is nonlinear, nondifferentiable and integer: in summary, it is complex. In such a case, the use of exact solution methods would be excessively timeconsuming, in particular for practical purposes. The proposed TS is tested using the daily closing prices from January 2, 2001, to June 30, 2016, of eight Italian stocks of different economic sectors. As benchmark, we consider the same TS but with standard time-window lengths. Irrespective of their signs, both in-sample and out-of-sample performances achieved by the TS with optimized parameters are better than those achieved by the benchmark, highlighting that parameter optimization can play an important role in TA-based TSs.
In this paper we consider a simple trading system (TS) based on a set of Technical Analysis (TA) indicators. Their peculiarity is the dependence on the time window widths used to calculate them. To attempt to improve the performances of the TS, we optimize these parameters (that is the time-window widths) by the Particle Swarm Optimization (PSO), which is a metaheuristic used to solve global optimization problems. The use of PSO is necessary since the involved optimization problem is nonlinear, nondifferentiable and integer: in summary, it is complex. In such a case, the use of exact solution methods would be excessively time consuming, in particular for practical purposes. The proposed TS is tested using the daily closing prices from January 2, 2001, to June 30, 2016, of eight Italian stocks of different economic sectors. As benchmark, we consider the same TS but with standard time-window lengths. Irrespective of their signs, both in-sample and out-of-sample performances achieved by the TS with optimized parameters are better than those achieved by the benchmark, highlighting that parameter optimization can play an important role in TA-based TSs.
An evolutionary approach to improve a simple trading system
Marco Corazza
;Francesca Parpinel
;Claudio Pizzi
2017-01-01
Abstract
In this paper we consider a simple trading system (TS) based on a set of Technical Analysis (TA) indicators. Their peculiarity is the dependence on the time window widths used to calculate them. To attempt to improve the performances of the TS, we optimize these parameters (that is the time-window widths) by the Particle Swarm Optimization (PSO), which is a metaheuristic used to solve global optimization problems. The use of PSO is necessary since the involved optimization problem is nonlinear, nondifferentiable and integer: in summary, it is complex. In such a case, the use of exact solution methods would be excessively time consuming, in particular for practical purposes. The proposed TS is tested using the daily closing prices from January 2, 2001, to June 30, 2016, of eight Italian stocks of different economic sectors. As benchmark, we consider the same TS but with standard time-window lengths. Irrespective of their signs, both in-sample and out-of-sample performances achieved by the TS with optimized parameters are better than those achieved by the benchmark, highlighting that parameter optimization can play an important role in TA-based TSs.
| File | Dimensione | Formato | |
|---|---|---|---|
|
2017-Corazza_Parpinel_Pizzi-An_evolutionary_approach_to_improve_a_simple_trading_system-BOOK.pdf
non disponibili
Descrizione: Articolo nella versione dell'editore.
Tipologia:
Versione dell'editore
Licenza:
Accesso chiuso-personale
Dimensione
1.57 MB
Formato
Adobe PDF
|
1.57 MB | Adobe PDF | Visualizza/Apri |
I documenti in ARCA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
