Time series analysis with Python

Bc. Illia Prazdnyk

Master's thesis

Abstract:

This thesis is focused on analysis of time ordered data using Python. The theoretical part uncovers a bit about how people achieved the understanding of time that we have nowadays, starting from the ancient invention of sundials and ending on the conventions for standardizing international time zones. After this part, there is a big block dedicated to explaining the statistical and mathematical foundations on which the time series analysis is based, including the description of basic principles of two statistical models: ARIMA and VAR. After that there is an introduction to Python programming language and deep learning together with a presentation of LSTM model and how it works. The practical part presents possible use case for each time series forecasting model presented above using Python and its packages commonly used for data analysis and visualisation. These demonstrations work primarily with close price of publicly traded stocks. The resulting code is executed on a cloud server provided by Google, and the notebook with all the code is linked in the appendix.

Keywords

Time series, Python, Statistics, Machine learning

Language used: English

Date on which the thesis was submitted / produced: 31. 3. 2022

Thesis defence

• Supervisor: Ing. Petr Hanzlík, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme / field:
Systems Engineering and Informatics / Informatics