This dissertation was written as a part of the MSc in Data Science at the International Hellenic
University. As Machine Learning establishes its way through every sector of our lives over the
last decades, its utilization on Time Series analysis has been crucial. Machine Learning for
forecasting the future behavior depending on past time series data, is being used widely in
Economics, Science, Medicine and Business sectors. Transforming time series into networks
and applying graph theory and complex network theory on it, is a method used increasingly
over the two past decades, since it combines the extensive research on these fields.
The goal of this dissertation is to provide a comparative review of the time series models that
are widely used and also present network-based time series analysis methods commonly used
in the research field. The practical part of this paper uses a time series that consists of hourly
power price and demand data, for the area of Sardegna over the period of one year.
Collections
Show Collections
