<![CDATA[The volatile, dynamic fluctuations in cryptocurrency prices make it difficult to predict closing prices in time-series modeling. In addition to historical price data, the dynamics of the crypto market are also influenced by investor sentiment shaped by digital sources such as news and online media, reflecting their perceptions, expectations, and reactions to market conditions that impact price direction. This study develops a cryptocurrency closing price prediction model using a deep learning approach with Gated Recurrent Units (GRU) that integrates market price and sentiment data into a single integrated framework. The research stages include data collection via APIs, selection of relevant data, preprocessing such as normalization and the creation of time-series sequences, construction of the prediction model, and evaluation and visualization of results. The findings indicate that the model effectively captures the temporal patterns of price movements, generating predictions that align with the actual trends of Bitcoin and Ethereum. The inclusion of sentiment data provides supporting signals to represent market dynamics shaped by investor behavior. Thus, the integration of price and sentiment data yields predictions that are more accurate and better representative of the cryptocurrency market.]]>
Copyrights © 2026
