SUCIPTO, MUHAMMAD BAMBANG (2025) KOMPARASI MODEL STATISTIK DAN DEEP LEARNING: EVALUASI KINERJA ARIMA, EXPONENTIAL SMOOTHING, DAN LSTM UNTUK PREDIKSI HARGA TELUR AYAM RAS DI PROVINSI JAWA TENGAH. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (543kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (39kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (233kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (208kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (494kB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (24kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (125kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (548kB) |
Abstract
The fluctuation in chicken egg prices is a challenge for stakeholders in Central Java, requiring accurate forecasting methods to make informed decisions. This study compares the performance of statistical and deep learning models—ARIMA, Exponential Smoothing (ES), and LSTM—for predicting chicken egg prices. The research aims to determine the most suitable model for capturing price trends and reducing forecast errors. Using historical price data, the evaluation was conducted by measuring MAPE, MAE, RMSE, and R² metrics. The results demonstrate that ARIMA performs well, with a MAPE of 1.618, MAE of 379.382, RMSE of 510.571, and R² of 0.865, indicating low error rates and good variability explanation. However, LSTM outperforms ARIMA, achieving the best performance with a MAPE of 1.257786, MAE of 301.385149, RMSE of 443.300073, and R² of 0.923366, showcasing its ability to capture complex patterns. Conversely, ES performs poorly, with a MAPE of 5.709807, MAE of 1351.530246, RMSE of 1805.516867, and a negative R² (-0.301243), proving unsuitable for volatile data. This study highlights LSTM as the most effective model, followed by ARIMA for moderate accuracy, while ES is not recommended for such dynamic data.always holds in these matters to this principle of selection: he rejects pleasures to secure other greater pleasures, or else he endures pains to avoid worse pains. Keyword: Chicken egg price forecasting, ARIMA, LSTM, Exponential Smoothing, Statistical and deep learning models Fluktuasi harga telur ayam ras menjadi tantangan bagi para pemangku kepentingan di Jawa Tengah, sehingga diperlukan metode peramalan yang akurat untuk pengambilan keputusan yang tepat. Penelitian ini membandingkan kinerja model statistik dan deep learning—ARIMA, Exponential Smoothing (ES), dan LSTM—untuk memprediksi harga telur ayam ras. Penelitian ini bertujuan untuk menentukan model terbaik dalam menangkap tren harga dan mengurangi kesalahan peramalan. Dengan menggunakan data historis harga, evaluasi dilakukan berdasarkan metrik MAPE, MAE, RMSE, dan R². Hasil penelitian menunjukkan bahwa ARIMA memiliki kinerja yang baik dengan MAPE sebesar 1.618, MAE sebesar 379.382, RMSE sebesar 510.571, dan R² sebesar 0.865, yang menunjukkan tingkat kesalahan rendah dan kemampuan menjelaskan variabilitas data. Namun, LSTM melampaui ARIMA, mencatat performa terbaik dengan MAPE sebesar 1.257786, MAE sebesar 301.385149, RMSE sebesar 443.300073, dan R² sebesar 0.923366, menunjukkan kemampuannya dalam menangkap pola kompleks. Sebaliknya, ES menunjukkan performa buruk dengan MAPE sebesar 5.709807, MAE sebesar 1351.530246, RMSE sebesar 1805.516867, dan R² negatif (-0.301243), sehingga tidak cocok untuk data yang fluktuatif. Penelitian ini menyoroti LSTM sebagai model paling efektif, diikuti ARIMA untuk akurasi moderat, sementara ES tidak direkomendasikan untuk data yang dinamis. Kata kunci: Peramalan harga telur ayam, ARIMA, LSTM, Exponential Smoothing, Model statistik dan deep learning
Actions (login required)
 |
View Item |
