Cryptocurrency has the potential to reshape financial systems and introduce financial investments that are inclusive in nature, which has led to significant research in the prediction of cryptocurrency prices by employing artificial neural networks and machine learning models. Accurate short-term predictions are essential for optimizing investment strategies, minimizing risks, and ensuring market stability. Prior studies in time-series forecasting have successfully employed statistical methods like Auto-Regressive Integrated Moving Average (ARIMA) and machine learning algorithms such as Long Short-Term Memory (LSTM). The research results presented in this paper evaluate various statistical and machine learning algorithms, assessing their accuracy and effectiveness in modeling volatile cryptocurrency data for short-term forecasting. Additionally, the study explores diverse hyperparameter settings to enhance the performance of machine learning models. The highest performance is achieved by a hybrid model combining LSTM and Deep Neural Network (DNN), showcasing its effectiveness in forecasting cryptocurrency prices with improved accuracy and capability.