Deep Learning-Based Price Prediction of Solana Using BiLSTM Models and Adam Optimizer

Abstract

The rapid growth of cryptocurrency markets has attracted significant attention from investors and researchers due to their high volatility and complex price dynamics. Solana, as one of the prominent blockchain platforms, exhibits substantial price fluctuations that pose challenges for accurate forecasting using conventional statistical methods. These characteristics motivate the application of advanced data-driven approaches capable of modeling nonlinear and temporal dependencies in cryptocurrency price movements. This study proposes a deep learning-based framework for Solana price prediction using a Bidirectional Long Short-Term Memory (BiLSTM) network optimized with the Adam algorithm. The bidirectional architecture enables the model to capture contextual information from both past and future time steps, while the Adam optimizer enhances training stability and convergence efficiency. Historical Solana price data are preprocessed and transformed into supervised learning sequences before being used to train the proposed model. The main contribution of this research lies in the application and evaluation of the BiLSTM–Adam model for Solana price forecasting, providing insights into its effectiveness in handling highly volatile cryptocurrency time series. Model performance is evaluated using standard regression metrics, including Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). Experimental results demonstrate that the proposed approach achieves accurate predictions and closely follows actual price trends during the testing period. Future work may explore the integration of additional features, such as technical indicators and market sentiment data, as well as the application of advanced architectures for multi-step and long-horizon cryptocurrency price forecasting.