Rahayu, Santi Dewi (2021) Nowcasting Indeks Harga Konsumen Harian Menggunakan Dynamic Factor Model dan Support Vector Regression. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211950017006-Master_Thesis.pdf - Accepted Version Download (6MB) | Preview |
Abstract
Indeks Harga Konsumen (IHK) merupakan salah satu indikator ekonomi yang digunakan untuk mengukur inflasi. Badan Pusat Statistik (BPS) mempublikasikan IHK dan inflasi dalam frekuensi bulanan dengan time lag satu hari kerja. Kebijakan yang diambil berdasarkan nilai inflasi bulanan bisa jadi kehilangan momentum karena peristiwa yang terkait dengan inflasi sudah terjadi jauh hari sebelum inflasi atau IHK dipublikasikan. Oleh karena itu, perhitungan IHK harian diperlukan untuk menggambarkan perubahan harga mendekati realtime. Nowcasting dapat mengatasi masalah ini dengan memprediksi inflasi harian melalui prediksi IHK harian. Penghitungan IHK harian dilakukan dengan memasukkan data harga bahan pokok harian di Sistem Informasi Ketersediaan dan Perkembangan Harga Bahan Pokok (SISKAPERBAPO) Provinsi Jawa Timur, data harian Jakarta Interbank Spot Dollar Rate (JISDOR) Bank Indonesia dan data harian harga minyak mentah Brent berjangka dari Id Investing ke dalam model nowcasting kemudian divalidasi dengan IHK bulanan BPS. Metode nowcasting yang digunakan dalam penelitian ini adalah Time Series Regression (TSR), Support Vector Regression (SVR) dan Dynamic Factor Model (DFM) untuk memperoleh nowcasting IHK harian Provinsi Jawa Timur. Akurasi model IHK ditingkatkan dengan membandingkan DFM dengan TSR dan SVR berdasarkan nilai Root Mean Square Error (RMSE), symmetric Mean Absolute Percentage Error (sMAPE) dan Mean Absolute Deviation (MAD). SVR Polynomial merupakan metode terbaik untuk nowcasting IHK harian. IHK harian hasil nowcasting dari model IHK bulanan digunakan sebagai variabel respon pada model IHK harian. SVR- Polynomial merupakan metode terbaik untuk nowcasting IHK harian pada model IHK harian.
=====================================================================================================
Consumer Price Index (CPI) is one of the economic indicators used to measure inflation. Badan Pusat Statistik (BPS) publishes a monthly CPI and inflation with a time lag of one working day. Policies based on the monthly inflation rate could be losing momentum as the events associated with inflation had occurred long before inflation or CPI was published. Therefore, it is necessary to calculate daily CPI to describe near real-time price changes. Nowcasting can overcome this issue by predicting daily inflation through predicting daily CPI. The calculation of daily CPI is done by entering daily data price of basic commodities in Sistem Informasi Ketersediaan dan Perkembangan Harga Bahan Pokok (SISKAPERBAPO), daily Jakarta Interbank Spot Dollar Rate (JISDOR) from Bank Indonesia, and daily Brent crude oil futures prices from Id Investing into a nowcasting model and validated by monthly CPI published by BPS. The nowcasting method used in this study is the Time Series Regression (TSR), Support Vector Regression (SVR), and Dynamic Factor Model (DFM) applied to predict daily CPI nowcasts in East Java Province. The performance comparison between DFM, TSR, and SVR is evaluated based on the Root Mean Square Error (RMSE), symmetric Mean Absolute Percentage Error (sMAPE), and Mean Absolute Deviation (MAD). SVR Polynomial is the best method for daily CPI prediction. The daily CPI nowcasting results from the monthly CPI model are used as response variables in the daily CPI model. SVR- Polynomial is the best method for nowcasting daily CPI on the daily CPI model.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Nowcasting, Dynamic Factor Model, Support Vector Regression, Time Series Regression, Indeks Harga Konsumen, JISDOR, SISKAPERBAPO, RMSE, sMAPE, MAD, Nowcasting, Dynamic Factor Model, Support Vector Regression, Time Series Regression, Consumer Price Index, JISDOR, SISKAPERBAPO, RMSE, sMAPE, MAD |
| Subjects: | Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis |
| Depositing User: | Santi Dewi Rahayu |
| Date Deposited: | 10 Sep 2021 11:48 |
| Last Modified: | 12 Jul 2024 02:15 |
| URI: | http://repository.its.ac.id/id/eprint/91966 |
Actions (login required)
 |
View Item |