A study on soft computing approach in weather forecasting

Weather forecasts based on temperature, wind speed and relative humidity are very important attributes in agriculture sector as well as many industries which largely depend on the weather condition. Therefore, having accurate weather forecasting information may allow farmers to make good decision on...

Full description

Saved in:
Bibliographic Details
Main Author: Yen, Wee Khun
Format: Thesis
Language:English
Published: 2010
Subjects:
Online Access:http://eprints.utm.my/id/eprint/11144/5/YenWeeKhunFSKSM2010.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Weather forecasts based on temperature, wind speed and relative humidity are very important attributes in agriculture sector as well as many industries which largely depend on the weather condition. Therefore, having accurate weather forecasting information may allow farmers to make good decision on managing their farm. Soft computing is an innovative approach to construct computationally intelligent systems that are supposed to processes humanlike expertise within a specific domain, adapt themselves and learn to do better in changing environments, and explain how they make decisions. The weather forecasting model based on soft computing is easy to implement and produces desirable forecasting result by training on the given dataset. The technique of soft computing such as BPNN, RBFN, PSONN and ANFIS are used in this study to test the performance in order to investigate which technique for weather forecasting is most effective and least of error. 720 hours of Johor Bahru weather data are used in this study in order to test their result of prediction based on MSE and RMSE. Besides, the experiment regarding the effect of different input nodes which applies tapped delay line method and different hidden nodes are also used to investigate whether previous data affects the performance. The result shows that ANFIS with input temperature, humidity, wind speed, weather condition(t), and weather condition(t-1) with the previous data will give the lowest MSE and RMSE, 7.0853e-3 and 8.4174e-2 consequence than other soft computing approaches.