Institutional Repository of Vilnius University
Laiko eilutes į vaizdus transformuojančių metodų taikymas anomalijų detekcijai /
Abstract
dc:descriptionThis thesis investigates the applicability of time series to image transformation methods in combination with deep learning‐based classification models – convolutional neural networks (CNNs) – for anomaly detection tasks involving imbalanced datasets. Since CNNs are typically used to analyze two‐dimensional data, it is proposed to transform time series into images using RP, GASF, GADF, MTF, and FIRTS methods. This enables leveraging pre‐trained image classifiers and the advantages of deep learning. The study experimentally compares the impact of different transformation methods on the accuracy of anomaly classification using several datasets and various evaluation metrics. The aim of this research is to contribute to a deeper understanding of how these methods perform when applied to imbalanced datasets, where anomalies belong to the minority class.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Vasaitytė, Rugilė,
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:ELABAETD239032667&prefLang=en_US