Using machine learning for healthcare challenges and opportunities

Abstract

Machine learning (ML) and its applications in healthcare have gained a lot of attention. When enhanced computational power is combined with big data, there is an opportunity to use ML algorithms to improve health care. Supervised learning is the type of ML that can be implemented to predict labeled data based on algorithms such as linear or logistic regression, support vector machine, decision tree, LASSO regression, K Nearest Neighbor, and Naive Bayes classifier. Unsupervised ML models can identify data patterns in datasets that do not contain information about the outcome. Such models can be used for fraud or anomaly detection. Examples of clinical applications of ML include the formulation of various clinical decision support systems. An important public health application of ML is the identification and prediction of populations at high risk for developing certain adverse health outcomes and the development of public health interventions targeted to these populations. Various concepts related to ML need to be integrated into the medical curriculum so that health professionals can effectively guide and interpret research in this area.