International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 186 - Number 62 |
Year of Publication: 2025 |
Authors: Kinza Muneer, Ubaida Fatima |
![]() |
Kinza Muneer, Ubaida Fatima . Cryptocurrencies Analytics with Machine Learning and Human-centered Explainable AI: Enhancing Decision-Making in Dynamic Market. International Journal of Computer Applications. 186, 62 ( Jan 2025), 52-67. DOI=10.5120/ijca2025924418
High volatility, common among often referred-to cryptocurrencies, is determined by numerous direct and indirect factors, and thus one of the main issues of such prediction. Such inherent instability most times causes investment uncertainty throughout the digital currency market. Looking at recent years, the forecasting of cryptocurrencies’ prices has received a high level of significance owing to their high levels of volatility. This study focuses on predicting the prices of three major cryptocurrencies: These cryptocurrencies include Bitcoin, Ethereum, and Litecoin. In this regard, three ML algorithms: LSTM, SVM, and RF as well as LSTM-RF ensemble were employed to enhance the predictive precision of the predicted cryptocurrencies. Comparing all the models investigated, the hybrid LSTM-RF model exhibited the highest accuracy in predicting the relevant performance indices and outperformed the other traditional models and other single machine learning methods. In addition, this study also adopts Explainable Artificial Intelligence (XAI) approaches to create AI-generated interpretable human-centric visualizations. This approach opens the insights to simple users as it helps investors make sound decisions depending on the results given by the model.