Comparative Study on Predicting Crude Palm Oil Prices Using Regression and Neural Network Models

Full Text PDF PDF
Author(s) Azme Khamis | Nursu’aidah Abd. Wahab
Pages 88-94
Volume 5
Issue 3
Date March, 2016
Keywords Multiple linear regression, neural network model, crude palm oil prices
Abstract

Palm oil has known as the important source of vegetables oils in the global market. Malaysia is the one of the major producer and exporters of palm oil. An accurate forecasting on crude palm oil (CPO) prices is considered significant to the oil palm business. This study was conducted to identify suitable model between Multiple Linear Regression (MLR) model and Artificial Neural Network (ANN) model on predicting Malaysia crude palm oil (CPO) prices. The Malaysia crude palm oil was predicted by three other Malaysia primary commodity prices which are natural rubber (NR) prices, black pepper (BP) prices and cocoa beans (CB) prices. The analysis use weekly data on the prices from Jan 2004 until Dec 2013. The methods are compared to obtain the best model for predicting crude palm oil price. It was found that, the value of in ANN model is higher than MLR model by 20.61%. The value of mean squared error (MSE) in ANN model also lower compared to MLR model. Therefore, ANN model is preferred to be used as alternative model in estimating crude palm oil (CPO) prices compared to MLR model.

< Back to March Issue