Peramalan Harga Solana dan Kontinuitasnya Menggunakan Recurrent Neural Network-Gated Recurrent Unit-Long Short Term Memory (RNN-GRU-LSTM)

Alfadhilah, Fikki Irsyad (2025) Peramalan Harga Solana dan Kontinuitasnya Menggunakan Recurrent Neural Network-Gated Recurrent Unit-Long Short Term Memory (RNN-GRU-LSTM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003211031-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Investasi pada aset high risk berupa cryptocurrency memiliki risiko yang tinggi dikarenakan harga cryptocurrency mengalami volatilitas yang tinggi. Solana merupakan salah satu aset cryptocurrency yang berada dalam lima posisi teratas dengan kapitalisasi pasar tertinggi. Dalam penelitian ini, peneliti mengusulkan metode peramalan harga Solana yang dilengkapi dengan analisis berbagai variabel yang dapat menentukan kontinuitas pada Solana yang memuat harga, kapitalisasi pasar, total nilai terkunci, dan volume perdagangan. Dibangun model peramalan harga Solana dan kontinuitasnya menggunakan Recurrent Neural Network-Gated Recurrent Unit-Long Short Term Memory (RNN-GRU-LSTM). Dengan pendekatan ini, tujuan utama penelitian adalah mengidentifikasi model terbaik untuk mendapatkan peramalan harga Solana secara akurat, mendapatkan hasil atau tingkat akurasi yang didapatkan setelah dilakukan peramalan harga Solana, serta melakukan analisis kemungkinan terbaik Solana dalam mempertahankan eksistensi atau kontinuitasnya. Model terbaik untuk peramalan harga Solana yaitu model RNN dengan 32 layer RNN, batch size 32 dengan nilai MAPE sebesar 3,06%. Model terbaik untuk peramalan kapitalisasi pasar Solana yaitu model RNN dengan 128 layer RNN dengan nilai MAPE sebesar 2,78%. Model terbaik untuk peramalan total nilai terkunci Solana yaitu model hybrid RNN-LSTM dengan kombinasi 32 layer RNN dan 32 layer LSTM dengan nilai MAPE sebesar 3,29%. Dari hasil peramalan berbagai variabel pada Solana, Solana mengalami penurunan pada harga dan kapitalisasi pasar namun persentase penurunannya tidak signifikan, dan mengalami peningkatan pada total nilai terkunci, sehingga dalam tiga bulan ke depan Solana akan tetap dapat mempertahankan eksistensi atau kontinuitasnya.
==================================================================================================================================
Investing in high-risk assets such as cryptocurrency carries significant risk due to the high volatility of cryptocurrency prices. Solana is one of the top five cryptocurrencies by market capitalization. In this study, the researcher proposes a forecasting method for Solana’s price, complemented by an analysis of various variables that may determine Solana’s continuity, including price, market capitalization, total value locked (TVL), and trading volume. A forecasting model for Solana’s price and its continuity is developed using a hybrid Recurrent Neural Network–Gated Recurrent Unit–Long Short-Term Memory (RNN–GRU–LSTM) approach. The primary goal of this study is to identify the most accurate model for forecasting Solana's price, evaluate the prediction accuracy, and analyze the most likely outcome regarding Solana's ability to maintain its existence or continuity. The best-performing model for Solana price forecasting is the RNN model with 32 neurons and a batch size of 32, achieving a MAPE of 3.06%. The best model for forecasting Solana’s market capitalization is also an RNN model with 128 neurons, producing a MAPE of 2.78%. For forecasting Solana’s total value locked, the best model is a hybrid RNN–LSTM model with a combination of 32 RNN neurons and 32 LSTM neurons, achieving a MAPE of 3.29%. Based on the forecasting results for the various Solana variables, Solana is expected to experience a slight decline in price and market capitalization, though the decrease is not significant, and an increase in total value locked. Therefore, over the next three months, Solana is likely to maintain its existence or continuity.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Investasi high risk assets, Cryptocurrency, Solana, Long-Short Term Memory, Mean Absolute Percentage Error.
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Fikki Irsyad Alfadhilah
Date Deposited:	01 Aug 2025 08:42
Last Modified:	01 Aug 2025 08:42
URI:	http://repository.its.ac.id/id/eprint/126178

Actions (login required)

View Item