A heuristic method for forecasting chaotic time series based on economic variables

Abstract

Time series is one of the most attractive and mysterious mathematical subjects. Weather temperature, rainfall, water flow volume of a river and other similar cases in meteorology are known and predictable time series; amount of load peak, electricity price and other similar cases in electrical engineering are considerable time series. Time series forecasting is highly taken into account in economy. Stocks price in stock exchange market, currency equivalent rate in such market as Forex, world price of petroleum, sugar, gas, gold and other key stuffs are best known time series. The discovery of chaos in economics time such as stock exchange is highly regarded by scholars of economics. In recent years, chaos has proven in many economic time series such as stock changes. Also, it has been proven that discovery of chaos will help to forecast time series by intelligent algorithms better than before. In this paper, by propose a new heuristic method inspired from chaotic characteristic of economic time series, forecasts this time series by means of artificial neural networks. In proposed method, output of chaotic function is used to help time series prediction well.