Institutional Repository of Vilnius University
Laiko eilučių prognozavimas kiekybiniam rizikos valdymui /
Abstract
dc:descriptionSuccessful risk management requires continuous data checking, which may not be sufficiently effective or correct if done manually. This problem can be addressed by using time series forecasting models to predict future data values, thereby predicting the normal behavior of the data, so that unpredicted data behavior can be captured as anomalies. The paper analyses autoregressive integrated moving average models and applies them to risk management data to predict future data. The paper presents a method for improving the accuracy of forecasts by fitting different variations of data to the models and using additional data analysis functions and algorithms, the outputs of which are passed to the SARIMAX model as external variables. The results show that the most accurate forecasts are obtained when the SARIMAX model is trained on a standardized dataset, where the data is subjected to Fourier analysis and the output of the analysis is used as external variables for model training. Most of the additional functions used improved the accuracy of the model predictions compared to the baseline model, and the method of model prediction refinement using additional analysis tools presented in this paper is considered to be effective and efficient.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bačianskaitė, Goda,
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:ELABAETD192845860&prefLang=en_US