Predictive Analytics in Healthcare: The Use of Machine Learning for Diagnoses

Currently with COVID19 pandemic, health becomes an important factor in our daily life. There are many diseases, some are treatable, some are lethal. Predictive analytics is a method to predict an outcome based on historical data and has been used in many sectors including healthcare. In healthcare, all data are kept in an Electronic Health Record, and therefore can be used to predict healthcare issues including diseases. Machine Learning is a method can be used to select and train the data to make a decision. Different methods in machine learning have been used to predict diseases such as heart and lung diseases, cancer and many more. This paper discusses different methods such as Support Vector Machine, Neural Network with Decision Trees, Random Forest, K-Nearest Neighbor, and variations of those algorithms that have been used for prediction in the field of health and healthcare. The different methods are compared in terms of the accuracy. In conclusion, the accuracy of the predictions is acceptable, however some are still low. The challenges are to improve the accuracy using different machine learning algorithms or hybrid algorithms.