Implementation Of The CNN-LSTM Combined Model For Gold Price Prediction In Gold Stock Investment

Aditya, Aflah (2023) Implementation Of The CNN-LSTM Combined Model For Gold Price Prediction In Gold Stock Investment. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05211942000001_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (3MB) | Request a copy

Abstract

The gold share price is an essential investment factor. Trends and fluctuations in the price of gold shares have become a source of investment risk because they are highly complex and influenced by numerous variables. If the price of gold drops significantly in the future, fluctuations in the gold share price will also have an effect on gold mining firms. Therefore, predictions of gold share behavior are required to aid and support the making of prudent investment policy decisions and potential risk mitigation. In order to obtain gold share price prediction results in this final project, historical data from a gold mining company is required, followed by a prediction process whose results can be used as an investment decision-supporting analytical factor. Combined models of convolutional neural network (CNN) and long short-term memory (LSTM) are used to forecast data. Through its memory function, the LSTM is used to analyze the relationships between time series data. An investment analysis report is used to make investment decisions. The rate of return on shares in the price earnings ratio (PER) method is included in the investment analysis report as an additional decision-support factor. This factor will be considered when classifying stocks (undervalued or overvalued) and making investment decisions (buy or sell). On undervalued stocks, the investment decision is to consider purchasing the stock, while on overvalued stocks, the investment decision is to consider selling the stock. On the basis of the training and testing results of the CNN-LSTM model, it can be concluded that the combined use of CNN and LSTM models has produced more accurate model evaluation metrics. The MAPE for validation data is 6.59 percent, while for testing it is 4.24 percent. This indicates that the model is capable of making accurate predictions and capturing data patterns and trends effectively.
=================================================================================================================================
Harga saham emas merupakan faktor investasi penting. Tren dan fluktuasi harga saham emas menjadi sumber risiko investasi karena sangat kompleks dan dipengaruhi oleh banyak variabel. Jika harga emas turun secara signifikan di masa mendatang, fluktuasi harga saham emas juga akan berpengaruh pada perusahaan tambang emas. Oleh karena itu, prediksi perilaku saham emas diperlukan untuk membantu dan mendukung pengambilan keputusan kebijakan investasi yang hati-hati dan potensi mitigasi risiko. Untuk mendapatkan hasil prediksi harga saham emas pada tugas akhir ini diperlukan data historis dari perusahaan pertambangan emas yang dilanjutkan dengan proses prediksi yang hasilnya dapat digunakan sebagai faktor analitis pendukung keputusan investasi. Model gabungan jaringan saraf convolutional (CNN) dan memori jangka pendek panjang (LSTM) digunakan untuk meramalkan data. Melalui fungsi memorinya, LSTM digunakan untuk menganalisis hubungan antar data deret waktu. Laporan analisis investasi digunakan untuk membuat keputusan investasi. Tingkat pengembalian saham dalam metode price earning ratio (PER) dimasukkan dalam laporan analisis investasi sebagai tambahan faktor pendukung keputusan. Faktor ini akan dipertimbangkan saat mengklasifikasikan saham (undervalued atau overvalued) dan membuat keputusan investasi (beli atau jual). Pada saham undervalued, keputusan investasinya adalah mempertimbangkan untuk membeli saham tersebut, sedangkan pada saham yang overvalued, keputusan investasinya adalah mempertimbangkan untuk menjual saham tersebut. Berdasarkan hasil pelatihan dan pengujian model CNN-LSTM, dapat disimpulkan bahwa penggunaan kombinasi model CNN dan LSTM telah menghasilkan metrik evaluasi model yang lebih akurat. MAPE untuk data validasi sebesar 6,59 persen, sedangkan untuk pengujian sebesar 4,24 persen. Hal ini menunjukkan bahwa model tersebut mampu membuat prediksi yang akurat dan menangkap pola dan tren data secara efektif

Item Type:	Thesis (Other)
Uncontrolled Keywords:	gold stock investment, time series, prediction, convolutional neural networks, long short-term memory, investasi saham emas, deret waktu, prediksi, jaringan saraf konvolusional, memori jangka pendek panjang
Subjects:	H Social Sciences > HG Finance > HG4529 Investment analysis Q Science > Q Science (General) > Q325.5 Machine learning. R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Aflah Aditya
Date Deposited:	18 Sep 2023 01:48
Last Modified:	18 Sep 2023 01:48
URI:	http://repository.its.ac.id/id/eprint/104323

Actions (login required)

View Item