On Testing Efficiency of Karachi Stock Exchange Using Computational Intelligence

This paper tests the efficiency of the Karachi stock market. The efficient market hypothesis suggests that the current price of an asset reflects all information that can be obtained from historical data. According to the proponents of this hypothesis, the best strategy in the absence of any predictive ability is to buy and hold. The paper compares this buy and hold strategy against a computational intelligence based trading strategy which predicts the price of an asset. The prediction is then used to identify the buying and selling points of the asset. The strategy is based on neural networks whose weights are optimized through particle swarm optimization. Both buy and hold and computational intelligence based strategies are tested on KSE100 index values for the period June 2004 to April 2007. The results show that the computational intelligence based strategy out performs the buy and hold strategy.