Peramalan Curah Hujan Menggunakan Metode Extreme Learning Machine

Simamora, Rich Juniadi Domitri (2020) Peramalan Curah Hujan Menggunakan Metode Extreme Learning Machine. Sarjana thesis, Universitas Brawijaya.

Abstract

Curah hujan adalah ketinggian air hujan yang terdapat dan terkumpul di tempat yang datar, tidak meresap, tidak menguap dan tidak mengalir. Informasi mengenai curah hujan sangat penting terutama dibidang pertanian dan sipil. Pada bidang pertanian informasi curah hujan digunakan untuk menentukan jenis tanaman yang akan ditanam sesuai dengan intensitas curah hujan, memprediksi awal musim tanam dalam kalender tanam untuk meminimalisir resiko penanaman. Pada bidang sipil, digunakan sebagai penentu standar rancang keteknikan dalam melakukan perencanaan bangunan pengendalian bencana banjir. Curah hujan di atas normal akan menimbulkan masalah bencana alam seperti banjir dan tanah longsor. Curah hujan adalah bagian dari elemen cuaca dan salah satu proses meteorologi yang cukup sulit untuk diprediksi. Maka dari itu, diperlukan peramalan curah hujan agar masyarakat dan pemerintah dapat melakukan tindakan pencegahan terhadap masalah yang ada. Proses peramalan terbagi menjadi beberapa proses yang antara lain normalisasi data, peramalan dengan algoritme Extreme Learning Machine, denormalisasi data dan hasil error dengan MAPE. Berdasarkan hasil pengujian menggunakan data curah hujan daerah Poncokusumo dengan rentang waktu tahun 2002 sampai 2015 diperoleh nilai MAPE terkecil sebesar 3.6852 %, dengan banyak fitur sebanyak 4, banyak neuron pada hidden layer sebanyak 2, persentase data training 80%.

English Abstract

Rainfall is the height of rain water that is found and collected in a flat, not absorbed, does not evaporate and does not flow. Information about rainfall is very important especially in agriculture and civil. In agriculture, rainfall information is used to determine the type of plants to be planted in accordance with the intensity of rainfall, predicting the start of the growing season in the planting calendar to minimize the risk of planting. In the civil field, it is used as a determinant of engineering design standards in planning flood disaster control buildings. Above normal rainfall will cause natural disasters such as floods and landslides. Rainfall is part of the weather element and one of the meteorological processes that is quite difficult to predict. Rainfall forecasting is needed so that the community and the government can take preventative measures against the existing problems. The forecasting process is divided into several processes which include data normalization, forecasting with the Extreme Learning Machine algorithm, data denormalization and the results of errors with MAPE. Based on the test results using rainfall data in the Poncokusumo area with a span of years 2002 to 2015 obtained the smallest MAPE value of 3.6852%, with as many features as 4, many neurons in the hidden layer as much as 2, the percentage of training data 90%.

Other obstract

-

Item Type:	Thesis (Sarjana)
Identification Number:	SKR/FILKOM/2020/15/052002977
Uncontrolled Keywords:	peramalan, curah hujan, Extreme Learning Machine, MAPE, forecasting, rainfall, Extreme Learning Machine, MAPE
Subjects:	000 Computer science, information and general works > 003 Systems > 003.2 Forecasting and forecasts
Divisions:	Fakultas Ilmu Komputer > Teknik Informatika
Depositing User:	Budi Wahyono Wahyono
Date Deposited:	10 Aug 2020 06:49
Last Modified:	04 Oct 2024 08:03
URI:	http://repository.ub.ac.id/id/eprint/180471

Text Rich Juniadi Domitri Simamora.pdf Download (5MB)

Actions (login required)

View Item