Cryptocurrency Forecasting: Unveiling the Future of Bitcoin Prices through Deep Neural Networks

Challa, Koushik Reddy (2023) Cryptocurrency Forecasting: Unveiling the Future of Bitcoin Prices through Deep Neural Networks. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (902kB) | Preview

Abstract

In recent years, the markets for cryptocurrencies, especially Bitcoin, have undergone significant growth and development. Therefore, it is now more important than ever to forecast future pricing movements accurately. Currently, there are forecasting methods not capable enough in the field of study that are extensive or innovative enough to effectively capture the temporal correlations and minute trends in Bitcoin price data. Our research aims to enhance prediction performance by utilizing Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) models, thereby addressing these issues. For this study, I aim to estimate the value of Bitcoin by employing time-series-based Neural Networks like RNNs and LSTM-based deep neural network models. I will then compare the outcomes with traditional machine learning models such as K-Nearest Neighbors (KNN), Linear regression, Ridge, and Support vector machine (SVM). The main objective of this project is to evaluate the performance of price predictions produced by deep neural networks compared to those generated by conventional machine learning models. The main aim of this study was to enhance the existing knowledge and techniques in accurately predicting the future value of cryptocurrencies. Furthermore, conduct comprehensive hyper-parameter tuning after finding that the deep neural network surpasses the ability to predict the future bitcoin prices from the traditional machine learning methods and ablation inquiry to showcase the effectiveness of the proposed deep neural network-based approach for both LSTM & RNN in the ultimate development of bitcoin future price prediction which can be advantageous to the investors to assist the well-driven decisions. From the results, we can demonstrate that deep learning models (LSTM-67, RNN-118) have shown significantly less MAE compared to traditional machine learning models.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Mulwa, Catherine UNSPECIFIED
Uncontrolled Keywords:	Bitcoin Prices; Deep Neural Networks; RNNs; LSTMs; Machine Learning; KNN; SVM; Linear Regression; Ridge
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HG Finance > Money > Digital currency Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	07 May 2025 11:54
Last Modified:	07 May 2025 11:54
URI:	https://norma.ncirl.ie/id/eprint/7503

Actions (login required)

View Item