What is Machine Learning? A Primer for the Epidemiologist.

Area of Science:

• Epidemiologic Sciences
• Computer Science
• Data Science

Background:

• Machine learning (ML) presents transformative potential for epidemiologic sciences, especially with the rise of Big Data.
• Classical epidemiologic methods may be insufficient for complex Big Data challenges.
• Bridging the technical and linguistic gap between ML and epidemiology is crucial for effective integration.

Purpose of the Study:

• To provide an overview of machine learning concepts and terminology for epidemiologists.
• To introduce common machine learning algorithms and ensemble methods.
• To summarize current epidemiologic applications of ML and recommend integration strategies.

Main Methods:

• Overview of core machine learning concepts and terminology.
• Introduction to 5 common ML algorithms (e.g., decision trees, support vector machines).
• Explanation of 4 ensemble-based ML approaches (e.g., random forests, gradient boosting).

Main Results:

• Summary of diverse epidemiologic applications of machine learning techniques found in published literature.
• Identification of key opportunities for ML in epidemiologic research.
• Discussion of challenges in integrating ML with existing epidemiologic methods.

Conclusions:

• Machine learning offers valuable tools for modern epidemiology, particularly for Big Data analysis.
• Addressing interdisciplinary barriers is key to successfully incorporating ML into epidemiologic research.
• Further research and collaboration can optimize the synergy between ML and epidemiology.