| Abstrak/Abstract |
The role of weather prediction is very critical because it can affect the transportation and agricultural activities mainly in agricultural country. In Yogyakarta, weather forecasting has been performed manually therefore automation using software should be designed to obtain a better accuracy and stability of prediction. The purpose of the research is designing a software using TXLAPS model (Tropical Extended Limited Area Prediction System)-based on intelligence system to be implemented in Yogyakarta. Intelligence system employed is the expert system algorithm through extracting knowledge, expertise and experience from the expert of weather prediction. Results of weather prediction software are tested to compare the accuracy of manual prediction with the accuracy of automatic prediction. The result of the software from the validation test, has ranged from 88.88 to 100% in which the minimum accuracy occurred in April 2010 and the maximum accuracy occurred in July 2009 and February 2010, while the average level of software accuracy is 95.14%. On the other hand, manual prediction has accuracy ranging from 75 to 100% in which the minimum accuracy occurred in September 2009 and the maximum accuracy occurred in February 2010, while the average accuracy rate is 90.74%. Thus, from the accuracy result show that the software can be applied in Yogyakarta to improve the accuracy of daily weather forecasting and to supersede the manual prediction. |
