Nanlohy, Yonlib Weldri Arnold (2017) Model Fungsi Transfer Multi Input Untuk Peramalan Curah Hujan Di Kota Surabaya. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
1315201021-Master_Thesis.pdf - Published Version Download (5MB) | Preview |
Abstract
Analisis data deret waktu pada dasarnya digunakan untuk melakukan analisis data yang mempertimbangkan pengaruh pengamatan periode sebelumnya. Analisis data deret waktu tidak hanya dapat dilakukan untuk satu variabel (univariate) tetapi juga dapat untuk banyak variabel (multivariat). Akhir-akhir ini peranan peramalan telah menjelajah ke berbagai bidang diantaranya bidang meteorologi yang menyangkut peramalan cuaca dan curah hujan. Metode yang sering kali digunakan untuk meramalkan curah hujan di kota Surabaya yaitu Autoregressive Integrated Moving Average (ARIMA). Model ARIMA merupakan model peramalan analisis data deret waktu tunggal atau model univariat. Model ARIMA hanya melihat ketergantungan curah hujan pada data masa lalu tanpa melibatkan ketergantungan curah hujan dengan variabel lain. Terdapat beberapa unsur cuaca (suhu, kelembaban udara, dan kecepatan angin) yang mempengruhi curah hujan sehingga perlu dilakukan peramalan dengan menggunakan model multivariat. Model fungsi transfer multi input merupakan model peramalan analisis deret waktu berganda (multiple) untuk model multivariat. Pada model fungsi transfer multi input terdapat deret output (y_t) yang diperkirakan akan dipengaruhi oleh deret input (x_t) dan input-input lain yang digabungkan dalam satu kelompok yang disebut deret noise (n_t). Tujuan dari penelitian ini yaitu untuk meramalkan curah hujan harian di kota Surabaya dengan model fungsi transfer multi input. Hasil peramalan dengan model fungsi transfer multi input akan dibandingkan dengan model ARIMA curah hujan harian di kota Surabaya untuk menilai keunggulannya. Data yang digunakan dalam penelitian ini yaitu data curah hujan, suhu, kelembaban dan kecepatan angin harian di kota Surabaya dari tahun 2008 sampai tahun 2015. Tingkat akurasi peramalan yang dihasilkan oleh setiap model peramalan diukur dengan kriteria Root Mean Square Error (RMSE). Hasil penelitian menunjukan bahwa model fungsi transfer multi input merupakan model peramalan terbaik pada peramalan curah hujan harian di kota Surabaya dengan nilai RMSE sebesar 11,6052 lebih kecil dibandingkan dengan model ARIMA sebesar 11,6056.
=================================================================
Analysis data of a time series is basically used to perform analysis data which consider the influence of the prior period of observation. Analysis data of time series not only can used for a single variable (univariate) but also to many variables (multivariate). Recently, the role of forecasting has been exploring into a various fields including the field of meteorology that concern to weather forecasting and rainfall. The method is often used to predict rainfall in the city of Surabaya, it is Autoregressive Integrated Moving Average (ARIMA). ARIMA models is forecasting model analysis data of single time series or univariate models. ARIMA models only look a dependence of rainfall on past data, without involving dependence on rainfall with other variables. There are several elements of weather (temperature, and humidity) that influenced a rainfall, so that needs to be done forecasting by using a multivariate model. In the multi input transfer function model there is an output series (y_t) that expected to be influenced by the input series (x_t) and the other inputs are combined into one group is called noise series (n_t). The purpose of this study is to forecast the daily rainfall in the city of Surabaya with a multi-input transfer function model. The result of forecasting with multi input input transfer model will be compared with ARIMA model of daily rainfall in Surabaya city to assess its superiority. Data used in this research are rainfall data, temperature, humidity and daily wind speed in Surabaya city from 2008 until 2015. The level of forecasting accuracy produced by each forecasting model is measured by Root Mean Square Error (RMSE) criteria. The results showed that the model of multi input transfer function is the best forecasting model in daily rainfall forecasting in the city of Surabaya with RMSE value is 11.6052 smaller than the ARIMA model is 11.6056.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | ARIMA, Multi Input Transfer Function, Rainfall. |
| Subjects: | G Geography. Anthropology. Recreation > G Geography (General) G Geography. Anthropology. Recreation > GA Mathematical geography. Cartography |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Yonlib Weldri Arnold Nanlohy |
| Date Deposited: | 31 Jan 2018 02:51 |
| Last Modified: | 08 Mar 2019 06:11 |
| URI: | http://repository.its.ac.id/id/eprint/45772 |
Actions (login required)
 |
View Item |