Long Short-Term Memory Recurrent Neural Network for Stroke Prediction
- Publication Type:
- Conference Proceeding
- Citation:
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2018, 10934 LNAI pp. 312 - 323
- Issue Date:
- 2018-01-01
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
© 2018, Springer International Publishing AG, part of Springer Nature. Electronic Healthcare Records (EHRs) describe the details about a patient’s physical and mental health, diagnosis, lab results, treatments or patient care plan and so forth. Currently, the International Classification of Diseases, 10th Revision or ICD-10 code is used for representing each patient record. The huge amount of information in these records provides insights about the diagnosis and prediction of various diseases. Various data mining techniques are used for the analysis of data deriving from these patient records. Recurrent Neural Network (RNN) is a powerful and widely used technique in machine learning and bioinformatics. This research aims at the investigation of RNN with Long Short-Term Memory (LSTM) hidden units. The empirical research is intended to evaluate the ability of LSTMs to recognize patterns in multi-label classification of cerebrovascular symptoms or stroke. First, we integrated ICD-10 code into health record, as well as other potential risk factors within EHRs into the pattern and model for prediction. Next, we modelled the effectiveness of LSTMs for prediction of stroke based on healthcare records. The results show several strong baselines that include accuracy, recall, and F1 measure score.
Please use this identifier to cite or link to this item: