Noverianto, Rizky Nanda (2018) Peramalan Penjualan Mobil Di Indonesia Dengan Model Hybrid Arimax - Deep learning neural network Dengan Hierarchy Time Series. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211440000062-Undergraduate_Theses.pdf - Published Version Download (3MB) | Preview |
Abstract
Total penjualan mobil di Indonesia saat ini terus meningkat. Dalam melakukan produksi mobil diperlukan perencanaan yang matang agar dapat meminimalkan pengeluaran dan memaksimalkan keuntungan. Penelitian ini bertujuan untuk mengetahui pemodelan dan peramalan jumlah penjualan mobil dan setiap merek dari Toyota, Honda, Daihatsu, Mitsubishi, dan Others yaitu gabungan semua merek selain 4 merek yang sudah disebutkan. Data yang digunakan dalam penelitian ini adalah data penjualan dan market share mulai tahun 2008 hingga 2017. Penelitian ini menggunakan dua tahap yaitu kajian simulasi dan terapan. Analisis data dilakukan dengan tiga metode yaitu ARIMAX, Deep Learning Neural Network (DLNN), dan Hibrida ARIMAX-DLNN. Hasil dari kajian simulasi menunjukkan bahwa penggunaaan metode DLNN cenderung lebih baik dari pada ARIMAX dan ARIMAX-DLNN baik pada noise linier dan nonlinier. Pada kajian terapan, hasil peramalan DLNN dibandingkan dengan ARIMAX dan ARIMAX-DLNN memberikan hasil yang lebih baik dalam meramalkan karena memiliki nilai Root Mean Square Error Prediction (RMSEP) dan Symmetric Mean Absolute Percentage Error Prediction (sMAPEP) yang lebih kecil yaitu dari 11 series sebanyak 9 series menggunakan metode DLNN dan 2 series menggunakan metode ARIMAX-DLNN.
=============================================================================================================================
The total of car market sales in Indonesia keeps increasing day by day. In the production of car, careful planning is needed in order to minimize expense and maximize profit. This research aims to find out the model and prediction of the amount of car sales hierarchily, from the brands of Toyota, Honda, Daihatsu, Mitsubishi, and Others, which are the combination of all brands except those four brands mentioned. The data used in this research are the sales and market share data starting from 2008 until 2017. Moreover, this research is done in two steps, which are simulation and practical studies. The data are analyzed by applying three methods, which are ARIMAX, Deep Learning Neural Network (DLNN), and ARIMAX DLNN Hybrid. The result from simulation study shows that the use of DLNN method tends to be better than ARIMAX and ARIMAX-DLNN either it is in linear or nonlinear noise. However, the result from ARIMAX-DLNN is way better compared to ARIMAX and DLNN in predicting the total of car sale in Indonesia in the scope of practical study because it has Root Mean Square Error Prediction (RMSEP) and Symmetric Mean Absolur Percentage Error Prediction (SMAPEP) values that is 9 out of 11 series better for using DLNN method, and 2 series better for using ARIMAX-DLNN method.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | ARIMAX, Deep learning neural network, Mobil, Market share, Hierarki, Variasi Kalender |
| Subjects: | Q Science Q Science > Q Science (General) > Q325.78 Back propagation T Technology > T Technology (General) > T174 Technological forecasting T Technology > TS Manufactures T Technology > TS Manufactures > TS167 Costs, Industrial |
| Divisions: | Faculty of Mathematics, Computation, and Data Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Rizky Nanda Noverianto |
| Date Deposited: | 05 Aug 2021 23:30 |
| Last Modified: | 13 Nov 2024 02:33 |
| URI: | http://repository.its.ac.id/id/eprint/56110 |
Actions (login required)
 |
View Item |