Analýza časových řad

Abstract

The aim of this master’s thesis was to describe and test several methods for time series analysis. Thesis will work with a dataset representing hourly natural gas consumption from 2013 to 2019. A significant part of the analysis will be based on time series decomposition, which will be used for describing the properties of the series and anomaly detection. Both selected and implemented methods for prediction fall into the category of machine learning. The programming language used for the implementation was Python.

Description

Subject(s)

time series, analysis, decomposition, prediction, machine learning, deep learning

Citation