Qi, Jing 
ORCID: 0000-0003-2476-7620, Burnside, Girvan ORCID: 0000-0001-7398-1346 and Coenen, Frans ORCID: 0000-0003-1026-6649
(2022)
Pathology Data Prioritisation: A Study Using Multi-variate Time Series.
.
|
Text
jingQi_DaWak_2022.pdf - Author Accepted Manuscript Download (438kB) | Preview |
Abstract
A mechanism to support the prioritisation of multi-variate pathology data, in the absence of a ground truth prioritisation, is presented. The motivation is the ever increasing quantity of pathology data that clinicians are expected to consider. The fundamental idea, given a previously unseen pathology result and the associated pathology history, is to use a deep learning model to predict future pathology results and then use the prediction to classify the new pathology result according to a pre-defined set of prioritisation levels. A further challenge is that patient pathology history, expressed as a multi-variate time series, tends to be irregularly time stamped and of variable length. The proposed approach used a Recurrent Neural Network to make predictions and a bounding box technique for the classification. The approach was evaluated using Urea and Electrolytes pathology data. The operation of the proposed approach was also compared with previously reported approaches, and was found to outperform these previous approaches.
| Item Type: | Conference or Workshop Item (Unspecified) |
|---|---|
| Uncontrolled Keywords: | Data ranking, Multivariate time series, Deep learning, Pathology data |
| Divisions: | Faculty of Health and Life Sciences Faculty of Science and Engineering > School of Electrical Engineering, Electronics and Computer Science Faculty of Health and Life Sciences > Institute of Population Health |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 17 Oct 2022 08:44 |
| Last Modified: | 26 Jul 2023 01:30 |
| DOI: | 10.1007/978-3-031-12670-3_13 |
| Related URLs: | |
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3165509 |
Dimensions
Dimensions
