Optimasi Peramalan Harga Saham Perusahaan Goto Berdasarkan Hasil Sentimen Pasar Menggunakan Distilbert Dan Gru (Gated Recurrent Unit)

Diqrelah, Wisnu Sevatiyan (2024) Optimasi Peramalan Harga Saham Perusahaan Goto Berdasarkan Hasil Sentimen Pasar Menggunakan Distilbert Dan Gru (Gated Recurrent Unit). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043201035_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (1MB) | Request a copy

Abstract

Perdagangan saham merupakan bentuk investasi menarik di pasar keuangan global, dan analisis harga saham serta peramalan pergerakan harga saham menjadi penting dalam membuat keputusan investasi yang tepat. Penelitian ini penting dalam menggabungkan teknologi pemrosesan bahasa alami DistilBERT dengan metode Gated Recurrent Unit (GRU) untuk analisis sentimen dan peramalan harga saham dalam pasar saham yang dinamis. Pada hasil DistillBERT, sentimen pasar diklasifikasikan menjadi tiga kategori sebagai prediktor: positif, negatif, dan netral. Metode ini diharapkan bermanfaat untuk menganalisis sentimen yang mempengaruhi pergerakan harga saham perusahaan GoTo. Sementara itu, GRU diharapkan mampu meramalkan harga saham dengan akurasi tinggi, mengandalkan pola-pola kompleks dalam data historis. Penelitian ini diharapkan mampu memberikan panduan kepada investor dan pelaku pasar dalam mengoptimalkan keputusan investasi, meminimalkan risiko, serta meraih keuntungan lebih baik dalam investasi saham GoTo. Penelitian ini menggunakan data sekunder dari website www.id.investing.com untuk mendapatkan data harga saham PT GoTo Gojek Tokopedia Tbk dari 17 April 2022 hingga 27 September 2023. Selain itu, data judul dan deskripsi berita yang berkaitan dengan GoTo diperoleh dari website berita www.investor.id selama periode yang sama. Hasil analisis menunjukkan bahwa model terbaik untuk meramalkan harga saham GoTo adalah model GRU yang dikombinasikan dengan sentimen pasar. Meskipun model tersebut mampu meramalkan pola harga secara tepat, terdapat selisih harga dengan harga saham asli. Oleh karena itu, investor GoTo disarankan untuk menjual kepemilikan saham pada investasi jangka pendek, sementara calon investor disarankan untuk menunggu harga mencapai titik terendah sebelum melakukan analisis lebih lanjut. Untuk penelitian selanjutnya, disarankan menambah kombinasi faktor, selain sentimen pasar, agar hasil peramalan lebih akurat dan presisi.
=================================================================================================================================
Stock trading is an attractive form of investment in the global financial market, and stock price analysis and forecasting of stock price movements are important in making informed investment decisions. This research is important in combining DistilBERT natural language processing technology with the Gated Recurrent Unit (GRU) method for sentiment analysis and stock price forecasting in a dynamic stock market. In the DistillBERT results, market sentiment is classified into three categories as predictors: positive, negative, and neutral. This method is expected to be useful for analyzing the sentiment that affects the stock price movement of GoTo companies. Meanwhile, GRU is expected to be able to forecast stock prices with high accuracy, relying on complex patterns in historical data. This research is expected to provide guidance to investors and market participants in optimizing investment decisions, minimizing risks, and achieving better profits in GoTo stock investments. This study uses secondary data from the website www.id.investing.com to obtain data on the share price of PT GoTo Gojek Tokopedia Tbk from April 17, 2022 to September 27, 2023. In addition, data on news titles and descriptions related to GoTo were obtained from the news website www.investor.id during the same period. The analysis results show that the best model for forecasting GoTo stock prices is the GRU model combined with market sentiment. Although the model is able to forecast the price pattern correctly, there is a price difference with the original stock price. Therefore, GoTo investors are advised to sell stock holdings on shortterm investments, while potential investors are advised to wait for the price to bottom out before conducting further analysis. For future research, it is recommended to add a combination of factors, in addition to market sentiment, so that the forecasting results are more accurate and precise.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	DistilBERT, GRU, GoTo Stock Price Forecasting, Peramalan Harga Saham GoTo
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Wisnu sevatiyan diqrelah
Date Deposited:	26 Jul 2024 02:33
Last Modified:	26 Jul 2024 02:33
URI:	http://repository.its.ac.id/id/eprint/109021

Actions (login required)

View Item