PENERAPAN MODEL HIBRIDA CEEMDAN-LSTM PADA PREDIKSI HARGA KOMODITAS PANGAN STRATEGIS HORTIKULTURA DI INDONESIA

Kristo Josua Simangunsong, . (2026) PENERAPAN MODEL HIBRIDA CEEMDAN-LSTM PADA PREDIKSI HARGA KOMODITAS PANGAN STRATEGIS HORTIKULTURA DI INDONESIA. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf
Download (213kB)
[img] Text
AWAL.pdf
Download (619kB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only
Download (291kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only
Download (875kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only
Download (438kB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only
Download (1MB)
[img] Text
BAB 5.pdf
Download (243kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (270kB)
[img] Text
RIWAYAT HIDUP.pdf
Restricted to Repository UPNVJ Only
Download (120kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only
Download (2MB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only
Download (21MB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only
Download (690kB)

Abstract

This study proposes a hybrid forecasting model for the prices of strategic horticultural commodities, namely shallots, garlic, red chili peppers, and cayenne peppers, by integrating Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Long Short-Term Memory (LSTM). CEEMDAN decomposes the price series into several Intrinsic Mode Functions (IMFs) and one residual component, separating high-frequency irregular fluctuations from more stable low-frequency trends, thereby improving the quality of the input data. Each decomposed component is then independently predicted using an LSTM model before being reconstructed to obtain the final price forecast. The results demonstrate that the proposed CEEMDAN-LSTM model consistently outperforms the conventional LSTM model across all commodities, as indicated by lower MAE, RMSE, and MAPE values, along with higher R² values. Compared with the standard LSTM model, the proposed hybrid approach reduces MAE by 7.28%–48.37%, RMSE by 19.37%–52.02%, and MAPE by 4.17%–47.83%. Furthermore, the model is employed to generate 14-day-ahead price forecasts using a recursive prediction strategy with a 95% prediction interval to quantify forecasting uncertainty. The proposed approach provides valuable insights for farmers and policymakers while also serving as an effective decision-support tool for agribusiness.

Item Type: Thesis (Skripsi)
Additional Information: [No.Panggil: 2110511111] [Pembimbing 1: Nurul Chamidah] [Pembimbing 2: Nurul Afifah Arifuddin] [Penguji 1: Muhammad Adrezo] [Penguji 2: Kharisma Wiati Gusti]
Uncontrolled Keywords: horticultural commodities, price prediction, time series, CEEMDAN, LSTM
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Fakultas Ilmu Komputer > Program Studi Informatika (S1)
Depositing User: KRISTO JOSUA SIMANGUNSONG
Date Deposited: 30 Jul 2026 07:39
Last Modified: 30 Jul 2026 07:39
URI: http://repository.upnvj.ac.id/id/eprint/52175

Actions (login required)

View Item View Item