CRYPTOCURRENCY PRICES PREDICTION: A TIME SERIES ANALYSIS OF CRYPTOCURRENCY PRICES TO PREDICT FUTURE TRENDS AND INFORM TRADING STRATEGIES
- Project Research
- 1-5 Chapters
- Quantitative
- Simple Percentage
- Abstract : Available
- Table of Content: Available
- Reference Style: APA
- Recommended for : Student Researchers
- NGN 4000
ABSTRACT
The objective of this study is to enhance comprehension in the domain of cryptocurrency price forecasting by doing a thorough analysis of time series data. The primary goal is to assess the efficacy of different predictive models, encompassing both conventional statistical techniques and contemporary machine learning methodologies. The study centres on Bitcoin and Ethereum, which are prominent cryptocurrencies, and employs yearly price data from 2015 to 2023 to assess the precision of the forecast. The primary strategies employed are autoregressive integrated moving average (ARIMA) models, generalised autoregressive conditional heteroscedasticity (GARCH) models, and advanced machine learning techniques such long short-term memory (LSTM) networks. The study demonstrates that whereas ARIMA and GARCH models offer excellent insights into linear trends and volatility patterns, machine learning algorithms, particularly LSTMs, greatly enhance forecast accuracy by capturing nonlinear dynamics and intricate patterns. The research also emphasises the influence of external events, such as market volatility and regulatory changes, on projected outcomes. This study offers valuable insights for traders and investors seeking to manage unpredictable cryptocurrency markets by evaluating model performance using measures such as root mean square error (RMSE), mean absolute error (MAE), and R-Squared. The findings enhance our understanding of financial forecasting by showcasing the superior efficacy of machine learning methods in predicting bitcoin prices. The findings of this study emphasise the necessity of using sophisticated modelling approaches and considering external variables in order to enhance the precision of forecasts. Potential areas for future research involve investigating hybrid models and analysing the impact of extra external variables on prediction accuracy.
Keywords: Cryptocurrency, Price Prediction, Time Series Analysis, ARIMA, GARCH, Machine Learning, LSTM