Machine learning applied to healthcare: a conceptual review

Area of Science:

• Health Informatics
• Computer Science
• Artificial Intelligence

Background:

• Technological integration in healthcare is crucial for enhancing decision-making and clinical outcomes.
• Machine learning (ML) offers significant potential for advancing healthcare procedures.
• Understanding the prevalent ML algorithms in healthcare is essential for future research and application.

Purpose of the Study:

• To systematically review and identify the primary machine learning algorithms utilized in healthcare.
• To categorize these algorithms based on a comprehensive literature review.
• To provide a foundation for further research into ML applications in the medical field.

Main Methods:

• A systematic literature review was conducted over a twenty-year period.
• 173 studies were analyzed based on predefined inclusion criteria.
• Identified algorithms were grouped into distinct typologies.

Main Results:

• The review identified 173 relevant studies published within the last two decades.
• Machine learning algorithms were successfully categorized into three main typologies: Supervised Learning, Unsupervised Learning, and Deep Learning.
• A significant portion of the analyzed works (59 studies) fell under these identified ML groups.

Conclusions:

• Supervised Learning, Unsupervised Learning, and Deep Learning represent the dominant machine learning paradigms in current healthcare applications.
• This systematic review provides a clear overview of ML algorithm usage in healthcare, supporting future technological advancements.
• The findings are expected to encourage further exploration and implementation of machine learning in healthcare settings.