An Accurate Missing Data Prediction Method using LSTM based Deep Learning for Health care

In this paper, an accurate missing data prediction method using Long Short-Term Memory (LSTM) based deep learning for health care is proposed. Physiological signal monitoring, especially with missing data, is a challenging task in health-care monitoring. The reliable and accurate acquisition of many physiological signals can help doctors to identify and detect potential health risks. In general, the missing data problem arises due to patient movement, faulty kits, incorrect observation or interference of the network. Subsequently, this problem leads to poorly diagnosed results. The ability of LSTM model to learn long-term dependencies enables it for efficient missing data prediction. In this paper, we proposed two LSTM model for 5-step and 10-step prediction. The dataset used is MIT-BIH normal person ECG data. The experimental results obtained using the LSTM method outperforms the Linear Regression and Gaussian Process Regression (GPR) method.