Technical Analysis of Extended Kalman Filter in High Frequency Trading

Abstract

obtaining to the method with the least prediction error is one of the challenging issues of financial and investment markets analyzers. In high-frequency trading strategies predicting the most accurate price of the securities is the only way to benefit from the markets. This paper presents a new approach to set the functional parameters of the Kalman filter, dynamically according to the market environment changing with respect to the technical analysis indexes, the parameter Q is measured by simple artificial neural network.And feed to the EKF algorithm to predict the next price for the securities. The method is experimented on the securities of the Iran stock exchange market to show its desirability in short-term prediction.