{"title":"A R2N2 Approach For Cardiac Behavior Forecast on Non-Trending Big HealthCare Data","authors":"A. Haque, Tariq Mahmood, S. Ghani","doi":"10.1109/AECT47998.2020.9194156","DOIUrl":null,"url":null,"abstract":"Medical Science and Healthcare has made significant developments for the provision of better and effective cures of diseases to people. Specially the engagement of body worn devices generating electronic health record (EHR) has made patient’s condition analysis very convenient for consultants in realtime. Currently the usefulness of these EHR are subjective to understand the current situation of patient and apply treatment against that. However this massive amount of data can further be used for predictive and forecasted analytics which will allow before hand cure and patient condition information to medical institutions. Generally the EHR contains time components and can be used for time series analysis. Since the generation of EHR is high in velocity and volume so simple time series will not yield effective and accurate results. For the purpose we have used Residual Recurrent Neural Network (R2N2) instead of simple time series analysis in our research work for forecasting patient’s cardiac behavior. The novelty in our model is that our R2N2 is a composition of VARMAX and LSTM. The model works on an extrapolative approach and uses last result as an input for next value forecast with an accuracy of 92.7%. We compare our result and outcome with all possible related work and found that the accuracy of forecast is higher than others and the response is in near realtime which is the requirement of medical institution. Our work can be used for medical institutions and healthcare sectors under surveillance as a support to consultants for their practice on patients.","PeriodicalId":331415,"journal":{"name":"2019 International Conference on Advances in the Emerging Computing Technologies (AECT)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2020-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 International Conference on Advances in the Emerging Computing Technologies (AECT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/AECT47998.2020.9194156","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Medical Science and Healthcare has made significant developments for the provision of better and effective cures of diseases to people. Specially the engagement of body worn devices generating electronic health record (EHR) has made patient’s condition analysis very convenient for consultants in realtime. Currently the usefulness of these EHR are subjective to understand the current situation of patient and apply treatment against that. However this massive amount of data can further be used for predictive and forecasted analytics which will allow before hand cure and patient condition information to medical institutions. Generally the EHR contains time components and can be used for time series analysis. Since the generation of EHR is high in velocity and volume so simple time series will not yield effective and accurate results. For the purpose we have used Residual Recurrent Neural Network (R2N2) instead of simple time series analysis in our research work for forecasting patient’s cardiac behavior. The novelty in our model is that our R2N2 is a composition of VARMAX and LSTM. The model works on an extrapolative approach and uses last result as an input for next value forecast with an accuracy of 92.7%. We compare our result and outcome with all possible related work and found that the accuracy of forecast is higher than others and the response is in near realtime which is the requirement of medical institution. Our work can be used for medical institutions and healthcare sectors under surveillance as a support to consultants for their practice on patients.