THE 10–DAY RAINFALL PREDICTION MODEL IN INDRAMAYU DISTRICT : Abstract

• Main Author: <br> NIM. 224 03 010, Indragustari
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/6278
• Institution: Institut Teknologi Bandung

</i><b>Abstract: <i></b><p align="justify"> Rainfall clustering can overcome the unfavourable data problem. Rainfall cluster have to revised as the effect of land use, land cover, climate and weather changes. Rainfall prediction by using ANFIS and wavelet demonstrate good skill for rainfall prediction.<p align="justify"> In this research clustering of rainfall using smallest euclidian distance from some principal components that obtained from principal component analysis and the test of the rainfall prediction technique with wavelet transform and ANFIS has been applied to mean of ten days rainfall data each group with variation length of input data at district of Indramayu.<p align="justify"> Cluster analysis of rainfall yield 6 (six) rainfall group referred as season prediction area. Here in after, wavelet spectrum analysis indicate that rainfall in all group dominant influent by monsoon, while effect from El Niño, inter–tropical convergence zone and Madden–Julian oscillation occur with weak intensity, and rainfall prediction verification in each season prediction area from both of technique using value of RMSE and Pearson Correlation. This verification show that longer input length yield more accurate prediction, accuracy of prediction both of technique differ to every season. Combination both of technique can yield more accurate prediction.</p>