SOFT COMPUTING APPLICATION FOR RAINFALL PREDICTION IN KALIMANTAN
Clustering analysis of rainfall at Kalimantan using Competitive Neural Kohonen yields 5 groups area called Prediction Zone. Meanwhile, data spectrum shows that sunspot signal exist in time series of rainfall to all of the Prediction Zone with the biggest magnitude at Prediction Zone 2 and indicate...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/7575 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Clustering analysis of rainfall at Kalimantan using Competitive Neural Kohonen yields 5 groups area called Prediction Zone. Meanwhile, data spectrum shows that sunspot signal exist in time series of rainfall to all of the Prediction Zone with the biggest magnitude at Prediction Zone 2 and indicated that zone gives direct response to the sunspot phenomena. Role of the sunspot activity to the high cloud formation believed relationships to the cosmic rays flux that various at latitude. Monthly rainfall prediction with ANFIS Method and Neural Network done with 1 Predictor (rainfall) and 2 Predictors (combine between cosmic rays and sunspot) at various length of data that is 45 years, 30 years, and 15 years and 46 years data length for yearly prediction (2007-2020). Over all, output 1 Predictor ANFIS Method shows small average value RMSE (Root Mean Square Error) for monthly prediction. But, for yearly prediction, 2 Predictors ANFIS Method shows more accurates. That’s way, sunspot and cosmic rays phenomena as predictor needs to be considered for long term prediction because gives better accuracy then only using rainfall as predictor. |
|---|