Gold Price Prediction using Eight Neighborhood Non Linear Cellular Automata

Abstract

Gold price prediction is pronounced as one of the dynamic problems addressed by many researchers. Several researchers have proposed various methods to predict the gold price; still there is a potential room to a new novel method with more accuracy and precise prediction. This paper proposes a supervised classifier with an Eight Neighborhood -Non Linear Cellular Automata to predict the exact gold price. The input for this proposed classifier is taken from the time series data of the last ten years in India. The classifier is trained and tested to give daily predictions to the users, which helps many investors to decide the time to buy or sell gold. This classifier is trained to process large amount of statistical data, process the business decisions listed in news articles. Based on all these parameters it predicts the gold variations in future. The accuracy of the classifier tested with standard datasets was reported as 86.7%, which is considerably better when compared with the existing literature.