Genetic Algorithm for Trading Signal Generation Solution to trader’s dilemma: Is it right time to trade?
A supervised learning approach to generate trading signals (buy / sell) using a combination of technical indicators to detect rise and fall of the market with greater accuracy and generates realistic trading signals is proposed.
Abstract
— Technical indicators are widely used by traders and analysts in stock and commodity markets to predict market movements and to identify trading opportunity thereby enhancing trading profitability. There are more than 100 indicators used in practice today to understand the market behavior. Identification of the right combination of indicators for optimal portfolio performance has always been a challenging problem. In this paper, we propose a supervised learning approach to generate trading signals (buy / sell) using a combination of technical indicators. Proposed system uses a genetic algorithm along with principle component analysis to identify a subset of technical indicators which detect rise and fall of the market with greater accuracy and generates realistic trading signals. The performance of this new algorithm was tested using trading data obtained from National Stock Exchange (NSE), India. Simulation shows the enhanced profitability for the proposed trading strategy.