Institutional Repository of Vilnius University
Neišnešiotų naujagimių medicininių signalų klasifikavimas naudojant mažą duomenų kiekį /
Abstract
dc:descriptionIn this work we investigated medical patients classification problem. For the task we wanted to classify if preterm infants which have ductus arteriosus will recover or not and based on that decide if invasive treatment is necessary or not. In this problem data consisted of twelve hour long time series data which consisted of four metrics which described various metrics of patient at certain point of time. The dataset was small, and the classification classes were skewed. In this work similar data which wasn’t in the classification boundaries was used. The problem was approached with machine learning algorithms and data augmentation algorithms. In this work we have found that in case of limited data, using similar data which is outside of classification set gives us better results. In this work various of data augmentation parameters were tried and the best suiting one was found. Finally we have compared two different models alongside with various different data generation methods and we found that the highest classification is reached using LSTM model.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kryževičius, Mantas,
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:ELABAETD146241463&prefLang=en_US