Institutional Repository of Vilnius University
Žaliavų kainų indekso prognozė naudojant informaciją apie krizes /
Abstract
dc:descriptionThis study analyzes commodity market products price indices and their volatility during crises. As for many time series, which are using innovations, so for these ones as well we will be using modified ARIMA model - regression with ARIMA errors. This decision was made based on the fact that we have to indicate, when theses crisis are appearing and how time series should move according to the situation in the world. To be able to represent the crisis information we used a specific three word vocabulary, which was composed out of "karas", "recesija" and "epidemija" key words, search intensity values extracted from "Google Trends" database and these values we will be using as a regression part variable, while the actual time series of the commodity prices indices is used to model the error part. To be sure that our predictions will be done correctly the cross correlation will be used and after the application of this method we were sure that we selected the correct time series of commodity indices as the leading values. To apply the selected model we tested the stationarity of time series to make the forecast possible. For the components models are different, but the most common one to appear was regression with ARIMA($1,1,0$) errors and when compared with AR($1$) showed better results than standard one, indicating that the modified ARIMA has the better properties. To conclude we can say that the chosen "Google Trends" vocabulary has a positive effect to the forecast.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Belickas, Domas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD146237493&prefLang=en_US