LDWPSO based Bi-LSTM Model for Predicting the Missing Data in PHRs

In medical and health datasets, it's common for some information to be missing. When you look at datasets that are missing some information, you might get skewed results or less statistical power. Because of this, some data may be missing, which hurts the analysis of the data. These missing data should be replaced with appropriate values. This article suggests a way to estimate missing data using an optimised Bi-Directional Long Short-Term Memory (OBi-LSTM) applied to big data in healthcare. A memory gate is used in the suggested method to estimate missing data that comes up when collecting and processing data. In this study, information from the Korean National Health Nutrition Examination Survey (KNHNES) by the Korea Centers for Disease Control and Prevention (KCDC) is used. As typical South Korean health care records, they have a lot in common with personal health records (PHRs). On the basis of this, models can be made to estimate the missing data in PHRs. PHRs are also multimodal, which means that data can be collected from many different sources for a single object. The Bi-LSTM model is used in a multimodal setting because of this. By preprocessing, a KNHNES data set is made that has no missing values. In data set-based learning, a label is set as the first piece of data. The parameters of Bi-computational LSTM affect how fast the model works. Linearly decreasing weight particle swarm optimization is used in the research to deal with these problems (LDWPSO). There is a big difference in the number of parameters between the two models, but there isn't much difference in how accurate they are. A multi-modal approach can also cut down on the time it takes to process large amounts of data in places like hospitals and institutions.