Loading...
Improvement of features extraction of time series dataset by using autoencoder
Limpiyapirom, Narongchai
Limpiyapirom, Narongchai
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2020
Date Submitted
Keywords
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
In this research, we studied how to extract features vector from the Multivariate Time Series dataset (MTS) by using various types of Autoencoder. Do the classification method from these encoded vectors to show that the compressed data from Autoencoder still keeping enough essential features, although the size was reduced. The dataset that we used is blood measurements collected from patients after surgery; the result after measurement will represent that patients have surgical site infections or not. This dataset contains many missing data, so not just only deal with the compressing method, we are also facing with sparse data problem so that the imputation method will be performed to recover the data back. To sum up, our research will show the capability to compress features vector from sparse multivariate time series datasets while keeping enough information.
Associated Publications
Rights
Copyright of the original work is retained by the author.