Penerapan Model Hybrid ARIMAX-Quantile Regression untuk Peramalan Inflow dan Outflow Pecahan Uang Kartal di Jawa Timur

Salehah, Novi Ajeng (2017) Penerapan Model Hybrid ARIMAX-Quantile Regression untuk Peramalan Inflow dan Outflow Pecahan Uang Kartal di Jawa Timur. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1313100051 - Undergraduate_Theses.pdf - Published Version

Download (4MB) | Preview

Abstract

Bank Indonesia memiliki tujuan untuk memenuhi kebutuhan uang di masyarakat dan salah satu faktor yang menjadi pertimbangan dalam penyusunan Rencana Kebutuhan Uang oleh Bank Indonesia adalah nilai inflow dan outflow. Berdasarkan hal tersebut, peramalan inflow dan outflow menjadi hal penting demi mendukung tujuan Bank Indonesia dimana diperlukan suatu model peramalan yang mampu merekonstruksi pola tren, musiman dan variasi kalender serta heteroskedastisitas pada data inflow dan outflow. Penelitian ini menerapkan model hybrid ARIMAX-Quantile Regression serta beberapa model individu lainnya dengan menggunakan dummy efek variasi kalender Hari Raya Idul Fitri untuk meramalkan inflow dan outflow pecahan uang kartal di Jawa Timur. Data yang digunakan adalah data sekunder dari Bank Indonesia mengenai inflow dan outflow per pecahan pada periode Januari 2003 hingga Desember 2016. Terdapat empat model ARIMAX-Quantile Regression yang diterapkan pada penelitian ini. Hasil studi simulasi dan analisis pada data real menunjukkan bahwa model ARIMAX-Quantile Regression mampu menangkap pola variasi kalender pada data. Model ARIMAX-Quantile Regression baik digunakan untuk meramalkan outflow pecahan Rp50.000,00 dan Rp2.000,00 serta inflow pecahan Rp50.000,00 hingga Rp1.000,00 sedangkan pecahan lain baik diramalkan menggunakan model individu lain. Berdasarkan hasil perbandingan kriteria kebaikan model menggunakan RMSE, MAE dan MdAE ditunjukkan bahwa model hybrid ARIMAX-Quantile Regression menghasilkan ramalan titik yang baik dibanding metode individu yang lain. ================================================================================ One of the purposes of Bank Indonesia is to fulfill the needs of money in the community and some factors which considered to this situation are the value of inflow and outflow. Based on this condition, inflow and outflow forecasting become necessary to arrange the plan that will help Bank Indonesia to reach its purpose. A forecasting model is necessary to reconstruct trend, seasonal, and calendar variation patterns and also heteroscedasticity of inflow and outflow data. This study applies hybrid ARIMAX-Quantile Regression and other individual methods using dummy variable of Eid Mubarak calendar variation to predict amount of inflow and outflow. In this study, there are two main topics will be discussed, simulation study and case study in real data. There are four types of ARIMAX-Quantile Regression models that used. The results of this study show that ARIMAX-Quantile Regression model is able to capture the pattern of calendar variation in the data. The outflow data of Rp50.000,00 and Rp2.000,00 and the inflow of Rp50.000,00; Rp20.000,00; Rp10.000,00; Rp5.000,00; Rp2.000,00; and Rp1.000,00 are better forecasted using ARIMAX-Quantile Regression model than using time series regression or ARIMAX. In general, based on model fitting criteria comparison using RMSE, MAE and MdAE, ARIMAX-Quantile Regression method performs better forecast compared to other individual methods.

Item Type: Thesis (Undergraduate)
Additional Information: RSSt 519.535 Sal p
Uncontrolled Keywords: ARIMAX, Hybrid, Inflow, Outflow, Quantile Regression.
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA278.2 Regression Analysis
Divisions: Faculty of Mathematics and Science > Statistics > (S1) Undergraduate Theses
Depositing User: Novi Ajeng Salehah
Date Deposited: 07 Nov 2017 02:52
Last Modified: 05 Mar 2019 03:33
URI: http://repository.its.ac.id/id/eprint/48577

Actions (login required)

View Item View Item