Predictive Machine Learning Model to Predict the Price Movements of Cryptocurrency Meme Coin in the Solana Ecosystem

Abstract

The meme coin ecosystem on the Solana blockchain is showing rapid growth in 2024, thanks to its superior blockchain technology and strong community support. Meme coin projects such as BONK, and DOGWIFHAT have leveraged these advantages to thrive in the Solana ecosystem. This study aims to build a prediction model for the price movement of meme coin cryptocurrencies in the Solana ecosystem using the Long Short-Term Memory (LSTM) method, with Adam optimization. Historical meme coin price data is taken as the research dataset, and the model is trained using LSTM with several epoch variations to obtain the best results. The model is evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). The experimental results show that the LSTM model with Adam optimization can provide fairly accurate predictions, with the best performance at epoch 75 where the model successfully achieves a balance between training and testing data performance, without experiencing overfitting. This study provides valuable insights for investors, developers, and policymakers into the dynamics of the meme coin ecosystem on Solana and its potential use in the development of blockchain technology. With a better unders