Evaluation of LSTM and Quantum LSTM Approaches Applied in the Forecasting of Bitcoin Prices

Price predictions of the cryptocurrency market can be said to be difficult and risky because of the unique volatility this market possesses. Traditional models such as Long Short-Term Memory (LSTM) models often do capture temporal dependencies in time series data but those models tend to be limited in scalability and perform poorly in noisy settings. This research investigates Quantum Long Short-Term Memory (QLSTM) that is enhanced models by quantum computing technologies for higher efficiency and accuracy in a volatile ecosystem. Datasets of historical Bitcoin prices were created and both LSTM and QLSTM models were developed and validated in MATLAB environment. Their relative predictions were evaluated based on some parameters such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). The results indicate that although both models achieve the objective of predicting the price trends, the quantum model has higher prediction stability than the traditional model and is therefore reliable in volatile conditions. This reinforces the case for using QLSTM in predicting cryptocurrencies as a reliable tool for improving insight for both investors and traders. Following this, future work will focus on the development of quantum hardware platforms and corresponding algorithms that will significantly extend QLSTM applicability on even tougher prediction problems.