Prediksi Harga Crude Palm Oil (CPO) Menggunakan Algoritma Gated Recurrent Unit (GRU)

Murbiantoro, Muhammad Kevin Adnan (2023) Prediksi Harga Crude Palm Oil (CPO) Menggunakan Algoritma Gated Recurrent Unit (GRU). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06111640000045_Undergraduate_Thesis.pdf] Text
06111640000045_Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (1MB) | Request a copy

Abstract

Crude Palm Oil (CPO) menjadi salah satu olahan perkebunan kelapa sawit berupa minyak mentah kelapa sawit yang menjadi komoditi ekspor di Indonesia. Seiring waktu, kebutuhan CPO mengalami peningkatan karena banyaknya permintaan minyak mentah berbagai negara di dunia. Awal bulan Maret 2022, harga CPO mencetak rekor tertinggi yang mengakibatkan harga minyak goreng internasional melonjak tinggi, terutama negara Indonesia. Prediksi atau forecasting harga CPO dilakukan untuk memenuhi kebutuhan PT. X (bukan nama sebenarnya) dalam membeli bahan baku CPO, melalui analisis masalah fluktuasi harga CPO dengan trend data yang tidak pasti. Pada penelitian ini telah dilakukan penerapan metode Gated Recurrent Unit (GRU) dalam memprediksikan harga CPO. Metode GRU merupakan pengembangan dari metode Recurrent Neural Network (RNN) dimana metode ini dapat memprediksi harga CPO. Penelitian ini telah menggunakan data harian harga CPO mulai 10 Januari 2019 hingga 25 Mei 2023. Penelitian ini juga telah mengimplementasikan algoritma Gated Recurrent Unit Multiple Input-Multiple Output 1 Dimension (GRU_MIMO1D) dengan variasi metode yang menghasilkan hasil yang cukup baik. Namun metode GRU mendapatkan hasil yang lebih baik dibandingkan dengan metode GRU_MIMO1D pada data harga CPO, dengan nilai evaluasi performa metode GRU yaitu nilai Mean Absolute Error (MAE) sebesar 6.6916, nilai Mean Absolute Percentage Error (MAPE) sebesar 0.0127%, nilai Mean Square Error (MSE) sebesar 78.7476, dan nilai Root Mean Square Error (RMSE) sebesar 8.8739. Sedangkan nilai evaluasi performa metode GRU_MIMO1D yaitu nilai MAE sebesar 32.6568, nilai MAPE sebesar 2.9190%, nilai MSE sebesar 1976.5454, dan nilai RMSE sebesar 44.4583.
===============================================================================================================================
Crude Palm Oil (CPO) is one of the processed products of palm oil plantations, in the form of crude palm oil, which is an export commodity in Indonesia. Over time, the demand for CPO has increased due to the high demand for crude oil from various countries around the world. In early March of 2022, the CPO prices reached a record high, resulting in a significant increase in international cooking oil prices, especially in Indonesia. The prediction or forecasting of CPO prices is carried out to fulfill the needs of PT. X (not the actual name) in purchasing CPO raw materials, through an analysis of the problem of CPO price fluctuations with uncertain data trends. In this study, the Gated Recurrent Unit (GRU) method has been applied to predict CPO prices. The GRU method is a development of the Recurrent Neural Network (RNN) method, which is capable of predicting CPO prices. This study used daily CPO price data from January 10, 2019, to May 25, 2023. The study also implemented the Gated Recurrent Unit Multiple Input-Multiple Output 1 Dimension (GRU_MIMO1D) algorithm with various methods that produced reasonably good results. However, the GRU method obtained better results compared to the GRU_MIMO1D method on CPO price data, with evaluation performance values for the GRU method as follows: Mean Absolute Error (MAE) of 6.6916, Mean Absolute Percentage Error (MAPE) of 0.0127%, Mean Square Error (MSE) of 78.7476, and Root Mean Square Error (RMSE) of 8.8739. Meanwhile, the evaluation performance values for the GRU_MIMO1D method are: MAE of 32.6568, MAPE of 2.9190%, MSE of 1976.5454, and RMSE of 44.4583.

Item Type: Thesis (Other)
Uncontrolled Keywords: Crude Palm Oil (CPO), Forecasting, Gated Recurrent Unit (GRU), Time Series
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Q Science > QA Mathematics > QA280 Box-Jenkins forecasting
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Muhammad Kevin Adnan Murbiantoro
Date Deposited: 11 Sep 2023 03:16
Last Modified: 11 Sep 2023 03:16
URI: http://repository.its.ac.id/id/eprint/103946

Actions (login required)

View Item View Item