PENDUGAAN CURAH HUJAN DI WILAYAH SUMATRA DENGAN MENGGUNAKAN ANFIS

• 主要作者: RUMINTA
• 格式: Theses
• 語言: Indonesia
• 主題: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• 在線閱讀: https://digilib.itb.ac.id/gdl/view/5150
• 機構: Institut Teknologi Bandung

<b></i>Abstract :</b><i><p align=\"justify\"><br /> Rainfall has a very high variability, particularly in Indonesia, so that often became a big problem in water managements. As the most variabel climate\'s elements, rainfall must get more attention. In relation to its problem, a study on identificatin of rainfall model and prediction of rainfall has been conducted by using ANFIS in Sumatra\'s region.<br /> <p align=\"justify\">Identification of rainfall model used three ANFIS architecturs that respectively use one, two, and three inputs. ANFIS inputs are climate\'s elements that has very strong correlation to the rainfall. Rainfall models are tested for its fitting and are evaluated for its values of RMSE, MAD, and MAPE to get three the best rainfall model that represented pentad, decad, and month rainfall models. The best rainfall model is used to predict rainfall in Sumatra\'s region.<br /> <p align=\"justify\">The resuls of the ANFIS indentification indicated that the best rainfall model in one location to anothers in Sumatra\'s region had different inputs. The best pentad rainfall model in Jambi, Palembang, Pekanbaru, and Tanjung Karang used two inputs, but in Bengkulu and Medan used three inputs. Most inputs of the best pentad rainfall model is air humidity and rainfall. The best decad rainfall model in Sumatra\'s region used two inputs. Most inputs of the best decad rainfall model is air humidity and rainfall. The best month rainfall model in Bengkulu, Jambi, Palembang, and Tanjung Karang used two inputs, but in Medan and Pekanbaru used one input. Most inputs of the best month rainfall model is air temperature and rainfall.<br /> <p align=\"justify\">The results of ANFIS prediction indicated that rainfall in one location to anthers in Sumatra\'s region had different values of its minimum, maximum, average, and standard deviation and it\'s related to ANFIS\'s inputs.