Machine learning: a brief description and application in infectious disease research

Machine learning (ML) is an umbrella term that encompasses many algorithms used to help people understand large amounts of data. Over the last five years, ML has been increasingly used in many areas of life sciences, including health care. ML is based on the idea that machines need to be able to learn and adapt through experience, while artificial intelligence (AI) refers to a broader idea where machines can execute tasks "smartly". AI applies ML, deep learning and other methods to solve real problems in different fields of research, but general AI could be considered as a physical manifestation using ML to perform a task. For infectious diseases, the use of ML has great potential to help physicians make optimal clinical decisions by supporting diagnosis, prognosis, and selection of appropriate antimicrobial therapies, as well as to elucidate and understand the highly complex molecular mechanisms underlying these diseases. The authors offer a brief overview of this topic. Of the 596 publications on the usage of the ML in the research of infection published just in 2020, more than 50 publications on the use of ML as a research tool for infection diseases were reviewed. The main ML algorithms used in infectious disease research were identified. It could be concluded that ML has two main approaches - unsupervised and supervised learning and each has its own set of analytical tools. Most unsupervised learning techniques comprises cluster analysis and its algorithms fall into two main groups - hard clustering and soft clustering, whereas all supervised learning techniques are a form of classification and regression utilizing such an algorithms like logistic regression, support vector machine, neural network, decision tree, random forest, discriminant analysis and other analytical approaches. As a more objective ML approach in infectious disease research, supervised learning using appropriate algorithms is recommended.