PREDIKSI HARGA SAHAM MENGGUNAKAN ALGORITMA RECURRENT NEURAL NETWORK LONG SHORT-TERM MEMORY (LSTM)

IZZUDDIN, SALMAN (2023) PREDIKSI HARGA SAHAM MENGGUNAKAN ALGORITMA RECURRENT NEURAL NETWORK LONG SHORT-TERM MEMORY (LSTM). S1 thesis, Universitas Mercu Buana.

	Text (COVER) 01 Cover.pdf Download (379kB)
	Text (ABSTRAK) 02 Abstrak.pdf Download (29kB)
	Text (BAB I) 03 Bab I.pdf Restricted to Registered users only Download (47kB)
	Text (BAB II) 04 Bab II.pdf Restricted to Registered users only Download (213kB)
	Text (BAB III) 05 Bab III.pdf Restricted to Registered users only Download (96kB)
	Text (BAB IV) 06 Bab IV.pdf Restricted to Registered users only Download (327kB)
	Text (BAB V) 07 Bab V.pdf Restricted to Registered users only Download (77kB)
	Text (DAFTAR PUSTAKA) 08 Daftar Pustaka.pdf Restricted to Registered users only Download (87kB)
	Text (LAMPIRAN) 09 Lampiran.pdf Restricted to Registered users only Download (1MB)

Abstract

Stock investment is increasingly popular among young people and adults in Indonesia with increasing enthusiasm and technological developments that facilitate access to the capital market. The COVID-19 pandemic has significantly affected the Indonesian stock market by causing share price fluctuations in various sectors. The use of Recurrent Neural Network (RNN) algorithms such as LongShort Term Memory (LSTM) helps investors to predict stock price movements more accurately and practically. This research was conducted to find out how well the LSTM algorithm performs and to what extent the pandemic has impacted the accuracy of the LSTM model in predicting stock prices. The dataset used in this study is stock records from 20 Indonesian issuers in 5 different industries, which are divided into 2 categories, namely before pandemic and post pandemic. The method used to test the accuracy of the LSTM model predictions is the Wilcoxon Signed-Rank Test. The results of this study indicate that there are 11 predictions with significant value in the pre-pandemic stock dataset, and 6 predictions with significant value in the post-pandemic stock dataset. Keywords: Recurrent Neural Network, Long Short-term Memory, Wilcoxon Signed-rank Test, Public Issuer. Investasi saham semakin populer di kalangan anak muda dan orang dewasa di Indonesia dengan meningkatnya antusiasme dan perkembangan teknologi yang memudahkan akses pasar modal. Pandemi COVID-19 mempengaruhi pasar saham Indonesia secara signifikan dengan menyebabkan fluktuasi harga saham di berbagai sektor. Penggunaan algoritma Recurrent Neural Network (RNN) seperti LongShort Term Memory (LSTM) membantu investor untuk memprediksi pergerakan harga saham dengan lebih akurat dan praktis. Penelitian ini dilakukan untuk mengetahui seberapa baik performa algoritma LSTM dan sejauh mana pandemi berdampak pada keakuratan model LSTM dalam memprediksi harga saham.Dataset yang digunakan pada studi ini adalah record saham dari 20 emiten Indonesia di 5 industri berbeda, yang dibagi menjadi 2 kategori, yaitu sebelum pandemi dan pasca pandemi. Metode yang digunakan untuk menguji keakuratan dari prediksi model LSTM adalah Wilcoxon Signed-Rank Test. Hasil dari penelitian ini menunjukkan bahwa terdapat 11 prediksi dengan nilai signifikan pada dataset saham sebelum pandemi, dan 6 prediksi dengan nilai signifikan pada dataset saham pasca pandemi. Kata Kunci: Recurrent Neural Network, Long Short-term Memory, Wilcoxon Signed-rank Test, Emiten Saham.

Actions (login required)

View Item