Machine Learning

Software Column

by Allen McLean, RN, MN, MSc, PhD(c)

Allen is currently a PhD student in Health Sciences at the University of Saskatchewan (Saskatoon) in the Computational Epidemiology and Public Health Informatics Lab. His research interests include the development of computer modeling and simulation software for addressing health systems challenges, chronic diseases and health inequities at the population level, as well as mobile technologies applied in long-term care facilities. Allen previously attended the University of Victoria earning an MN and MSc (Health Information Science) in a unique dual degree program for Nursing Informatics professionals. Allen has over 20 years’ experience in healthcare as an ultrasound technologist, clinical educator, team leader and community health RN.

COLUMN

Many people believe that healthcare in Canada is rapidly approaching a critical tipping point, myself included. Demographic (our aging population) and epidemiologic (the increase in chronic conditions) changes mean the demand for health services will soon exceed traditional sources of supply. Informatics, specifically the intelligent automation of some healthcare services is a potential solution in some circumstances. Recognizing nursing’s central and indispensable role in healthcare, and assuming we acquire the knowledge and skills necessary to contribute in meaningful ways – informatics nurses are ideally situated to positively influence these inevitable changes.

Machine learning (ML) is a discipline within artificial intelligence (AI), and is the science of getting computers to do something, without explicit programming. In the past decade, machine learning has given us self-driving cars, practical speech recognition, effective web search, and a vastly improved understanding of the human genome. Machine learning is so pervasive today that you probably use it dozens of times a day without knowing it. Many researchers also believe ML is the best strategy to make progress towards human-level (general) AI.

The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data for making better decisions. Machine learning allows for the analysis of massive quantities of data. Machine learning is an area of specialization of statistics crossed with computer science, most notably with such areas as computational statistics, scientific computation, data visualization and computational complexity, which sounds daunting, but most ML techniques are accessible if you have an interest in these areas. Research in machine learning is concerned largely with the analysis and development of algorithms to explore, discover, visualize and model structure in data as well as to make predictions and decisions based on that structure. Most important, is a creative imagination – turning big data into useful information relies more on the questions we ask, rather than the specifics of the code we use.

Machine learning algorithms are often categorized as supervised or unsupervised. There are others, and a good resource should you wish to explore further comes from the University of Waterloo (2018). Supervised machine learning algorithms can apply what has been learned in the past to new data using labeled examples to predict future events. Starting from the analysis of a known training dataset, the learning algorithm produces an algorithm for making predictions about output values. The system is able to provide targets for any new input after sufficient training. The learning algorithm can also compare its output with the correct, intended output and find errors in order to modify the model accordingly. In contrast, unsupervised machine learning algorithms are used when the information used to train is neither classified nor labeled. Unsupervised learning studies how systems can infer a function to describe a hidden structure from unlabeled data. The system is rarely able to figure out the correct output precisely, but it explores the data and can draw inferences from datasets to describe hidden structures from unlabeled data.

ML has been applied successfully in many areas of nursing and healthcare. As a personal example, my dissertation research involves constructing a ML program that can autonomously classify (by risk) diabetics and predict the incidence of foot ulcers. The goal is targeted care to reduce this outcome.