Predictive Modelling for Healthcare Decision-Making Using IoT with Machine Learning Models

Abstract

Many diseases could be deadly and untreatable in the current scenario if they aren’t identified early. There is a need for identifying such diseases at their primary stage to start appropriate treatments for better results. High demand rises for investigating detailed clinical data, report summary, and medical imaging. Within a short time, it should be done with promptness. Machine Learning (ML) methods are brought to the limelight because of their excellence in identifying patterns in observational data. Research on public health policy says healthcare has grasped IoT analytics and ML methods for its predictive accuracy. Hence, the self-operating machines are most trustable for making medical records, diagnosing diseases, and performing real-time monitoring for patients. To educate new paradigms, the researchers took data from accident Data and narratives. To check its results, they compare the performance of non-trained traditional logistic models with trained new models on tabular and narrative-based data. To face the data imbalance-based challenges, they used synthetic data augmentation techniques. This chapter mainly focuses on various ML algorithms, their significance in computational biology, and how the different approaches have been implemented in health sectors for decision-making were also discussed. In recent times, neural network-based Deep Learning (DL) methods perform remarkably in healthcare sectors. They take a smaller percentage of people from the subgroup and use data augmentation to foretell how many days the employees are off from labour. The outcomes show the significance of predictors and how it has improved the F1-score in mining employees’ check-in time, days spent away from home, and using augmentation techniques and strategies.