Perbandingan GSTAR Dan ARIMA Filter Kalman Dalam Perbaikan Hasil Prediksi Debit Air Sungai Brantas

Hamsyah, Ilham Fauzi (2015) Perbandingan GSTAR Dan ARIMA Filter Kalman Dalam Perbaikan Hasil Prediksi Debit Air Sungai Brantas. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
1211100043-Undergraduate Thesis.pdf - Accepted Version

Download (5MB)

Abstract

ARIMA Box-jenkins adalah salah satu metode time series yang biasa digunakan untuk melakukan analisis data dan peramalan. Dalam kehidupan sehari-hari, kita sering menemukan data yang mempunyai keterkaitan antar waktu dan keterkaitan antar lokasi. Data seperti ini disebut data spasial. Debit air Sungai mempunyai keadaan yang heterogen pada setiap waktu dan lokasi pengukuran yang dipengaruhi sifat acak alam, sehingga karakteristik debit air disetiap lokasi berbeda. Untuk mendapatkan prediksi yang mempunyai tingkat error yang kecil, maka akan dilakukan perbandingan dua model yaitu model Generalized Space Time Autoregressive (GSTAR) dan model Autoregressive Integrated Moving Average (ARIMA) Filter Kalman. Algoritma Filter Kalman akan diterapkan pada hasil ramalan Pemodelan ARIMA dengan pengambilan derajat polinomial kesatu, dua, dan tiga untuk memperbaiki prediksi 14 hari ke depan. Hasil akhir menujukan bahwa Filter Kalman mampu memperbaiki hasil ARIMA dan mempunyai tingkat error yang lebih kecil dibandingkan dengan GSTAR(31) inverse jarak, yang ditunjukan melalui hasil simulasi berupa grafik dan diperjelas dengan nilai MAPE yang lebih kecil. ================================================================================================== ARIMA Box-Jenkins is one method of time series which is used to perform data analysis and forecasting. In daily life, we often find data that have a relation between the time and relation between locations. Data such as these are called spatial data. River water discharge have a heterogeneous situation at any time and location measurements influenced the random of nature, so that the water discharge characteristics at each location. To get the predictions that has a small error rate, it will be the comparison of two models, namely models Generalized Space Time Autoregressive (GSTAR) and models Autoregressive Integrated Moving Average (ARIMA) Kalman Filter. Kalman Filter algorithm will be applied to the results forecast by the ARIMA modeling decision-degree polynomial 1-st, 2-nd, and 3-rd to improve the prediction of the next 14 days. The final result vector that is able to improve the results of ARIMA Kalman Filter and have a level of error that is smaller than the GSTAR (31) inverse distance, which is demonstrated through simulation results in the form of graphs and clarified with a smaller MAPE value.

Item Type: Thesis (Undergraduate)
Additional Information: RSMa 519.536 Ham p
Uncontrolled Keywords: ARIMA, ARIMA Filter Kalman, polinomial derajat, GSTAR
Subjects: H Social Sciences > HA Statistics
Divisions: Faculty of Mathematics and Science > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Yeni Anita Gonti
Date Deposited: 10 Feb 2020 07:14
Last Modified: 10 Feb 2020 07:18
URI: http://repository.its.ac.id/id/eprint/74788

Actions (login required)

View Item View Item