Peramalan curah hujan sebagai pendukung kalender tanam padi di Kabupaten Bojonegoro menggunakan Metode Arima, Support Vector Regression dan Genetic Algorithm-Support Vector Regression

Suci, Kiki Wulan (2017) Peramalan curah hujan sebagai pendukung kalender tanam padi di Kabupaten Bojonegoro menggunakan Metode Arima, Support Vector Regression dan Genetic Algorithm-Support Vector Regression. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
1313100086-Undergraduate-Theses.pdf
Restricted to Repository staff only

Download (3MB) | Request a copy

Abstract

Kabupaten Bojonegoro dikenal sebagai lumbung padi di Jawa Timur. Sebanyak 33,31% lahan Bojonegoro digunakan se-bagai lahan sawah. Pada beberapa tahun terakhir produksi padi Kabupaten Bojonegoro mengalami fluktuasi yang salah satu pe-nyebabnya adalah iklim ekstrem. Iklim ekstrem dapat meng-akibatkan efek musim kemarau yang panjang serta adanya hujan ekstrem yang mengakibatkan petani mengalami kerugian. Oleh ka-rena itu, keberhasilan produksi padi sangat bergantung pada in-formasi mengenai data curah hujan yang tersusun dalam kalender tanam. Dalam penelitian ini dilakukan peramalan curah hujan da-sarian di Pos Cawak dan Kedungadem menggunakan metode ARIMA, Support Vector Regression (SVR) dan Genetic Algorithm-SVR (GA-SVR). Berdasarkan RMSE dan SMAPE metode GA-SVR menghasilkan peramalan yang lebih akurat. Berdasarkan forecast 6 bulan selanjutnya akan dibuat kalender tanam padi. Hasil ka-lender tanam padi pada bulan Juli 2016- Desember 2016 me-nunjukkan kebutuhan air untuk penanaman padi sawah tidak dapat terpenuhi. Petani dapat mengganti padi dengan menanam pala-wija Jika tetap menanam padi, maka petani dan pemerintah harus memastikan tersedianya cadangan air. ====================================================================================== Bojonegoro District known as a granary in East Java. A total of 33.31% of the land in Bojonegoro is used as a wetland. In recent years the production of rice in Bojonegoro fluctuated, which one of them caused by extreme climate. Extreme climate can caused continuously dry season and extreme rainfall can caused farmers suffer losses. Therefore, the success of rice production is highly dependent on information of the rainfall data which arranged in a planting calendar. In this study, rainfall forecasting is done per 10 days in Cawak and Kedungadem Station using ARIMA method, Support Vector Regression (SVR) and Genetic Algorithm-SVR (GA-SVR). Based on RMSE and SMAPE values, GA-SVR method gave better forecast accuracy. Rice planting calendar will be made based on the forecasting result of the next 6 months. The result of rice planting calendar on July 2016 – December 2016 indicates that the water requirement for rice cultivation cannot be fulfilled. Farmers may substitute it with palawija which require less water, but if they keep doing the rice planting, the farmers and government must ensure the availability of water reserves.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: ARIMA; Curah Hujan; Genetic Algorithm; Support Vector Regression; Rainfall
Subjects: G Geography. Anthropology. Recreation > GE Environmental Sciences
Q Science > QA Mathematics > QA278.2 Regression Analysis
Divisions: Faculty of Mathematics and Science > Statistics > (S1) Undergraduate Theses
Depositing User: KIKI WULAN SUCI
Date Deposited: 10 Mar 2017 07:52
Last Modified: 19 Dec 2017 06:02
URI: http://repository.its.ac.id/id/eprint/3560

Actions (login required)

View Item View Item