Abstract:
The objective of this master project presents a palm oil price prediction system via the
Internet to forecast a price trend on the palm oil. In this paper, the data, which is collected from
Suratthani Oil Palm Research Center for 5 years, from 2008 to 2013. Herein, the Artificial
Neural Network is utilized to create a learning model with 20 hidden layers and 0.1 for learning
rate. From the experiment, the predicted accuracy comparison of created learning model was
considered, which was compared between the Artificial Neural Network, Linear Regression
Analysis, Support Vector Regression, and Radial Basis Function techniques. The results
showed that the mean square error (MSE) of learning model implemented by Radial Basis
Function, Linear Regression Analysis, Support Vector Regression and Artificial Neural
Network were 0.2958, 0.0541, 0.0536 and 0.0418, respectively. From the result, the Artificial
Neural Network had the lowest MSE, then it was implemented a learning model for a palm oil
price prediction system via the Internet.