Modeling and Optimization of Signals Using Machine Learning Techniques

2Role of AI in Mortality Prediction in Intensive Care Unit Patients

Prabhudutta Ray¹, Sachin Sharma^2*, Raj Rawal³ and Dharmesh Shah²

¹Institute of Advanced Research, Gandhinagar, Gujarat, India

²Indrashil University, Mehsana, Gujarat, India

³Gujarat Pulmonary and Critical Care Medicine, Ahmedabad, Gujarat, India

Abstract

Mortality rate and risk prediction in the intensive care unit (ICU) are two major concerns in hospitals. Prediction is highly dependent on the correct analysis of available data. Many data analysis techniques can provide sufficient information to correctly predict the survival of ICU patients during their admission time in the hospital. Despite methodological advances in machine learning algorithms, clinicians often use black box model. This flaw creates a difficulty in translating complex clinical parameters to corresponding model features. Regarding mortality prediction, it is found that different types of machine learning algorithms have different features that can be properly selected for the respective problem. This chapter focuses on the uses of the types of machine learning algorithms relevant to correct prediction related to survival of the patients.

Keywords: Intensive care unit (ICU), emergency departments (EDs), electronic medical records (EMR), machine learning (ML), sepsis-related organ failure assessment (SOFA), artificial intelligence (AI), early mortality prediction for intensive care unit (EMPICU), data mining (DM), directional coronary atherectomy ...