Peramalan Panjang Musim Hujan Menggunakan Adaptive Fuzzy Inference System
Abstract
Climate is a component of the ecosystem and an important natural factor which is dynamic and difficult to be controlled. The dynamic and diverse characteristics of climate need more accurate analysis to make climatic information be more useful in the agriculture sector. Uncertain climatic condition could be a limiting factor for crops productivity. Technically, in crops cultivation all climatic elements almost affect crops productivity and management. One of the solutions for this problem is by predicting the length of the rainy season to minimize the level of crops failure in agricultural product by using adaptive fuzzy inference system (ANFIS) method. This research used the Southern Oscillation Index (SOI) data and the onset data with the length of the rainy season of the previous year as the predictors to predict length of the rainy season of the current year. This research is focused at the Indramayu district. As the predictive model, we use adaptive neuro fuzzy system. Five models were trained for each of the five regions used, as well as one model for the average data of the five regions. The best prediction was obtained from the model on the fifth region. On this region, the correlation coefficient between the observed values and the predicted values is 0.69 and the RMSE value is 4.09 (measured in tens of days). From this result, we conclude that the performances of the models were still not satisfactory.
Collections
- UT - Computer Science [2322]