Nurse Developer Column

By Raymund John Ang, RN, MAN, PhD St

Raymund is currently a PhD student at the Holy Angel University (Philippines), and works as a clinical analyst in Pennsylvania, USA. He is the project lead of the Open Nursing Information System (Open-NIS) Project

Citation: Ang, R. J. (2023). Rule-based and Machine Learning Approaches to AI. Nurse Developer Column. Canadian Journal of Nursing Informatics, 18(2). https://cjni.net/journal/?p=11562

The term Artificial Intelligence (AI) has been a buzzword when it comes to how technological innovation affects patient care, and healthcare in general. From healthcare conversational agents (chatbots) to the application of machine learning algorithms to aid in diagnosis and clinical decision-making, there are a multitude of ways that advancement in technology can be applied to the healthcare field. But does it always involve the implementation of complex algorithms and consuming vast amounts of computing power? Or are decision trees sufficient to address some of the basic health issues?

Designing a system on a rule-based approach takes advantage of human expertise and knowledge, and the structure of a decision tree (otherwise known as expert systems). Logically, decision paths are based on a series of IF-THEN statements (Chan et al., 2022). On the other hand, machine learning systems utilize more complex algorithms and make use of a larger chunk of data set than rule-based systems. IF-THEN statements may not be sufficient, and instead, patterns are detected based on algorithms and the use of big data analytics (Hasan & Rizvi, 2022).

One might ask – which approach is better? Would employing machine learning be a more suitable approach since complicated algorithms and big data analytics are involved? Is rule-based or expert system approach less effective when it comes to developing clinical decision systems?

Having a mindset of rule-based versus machine learning for artificial intelligence does not take full advantage of what either has to offer. Instead, rule-based and machine learning strategies can be integrated into the workflow. An example is when mathematical or numerical results from machine learning calculations are built into conditional IF-ELSE statements of a rule-based system. In this case, the user benefits from both approaches. And when applied to healthcare delivery service, the patient can ultimately benefit from a holistic and integrative approach to how data is analyzed, and how information systems are developed. This adds an exciting avenue on which nurse informaticists could play an important role.

References

Chan, L., Hogaboam, L., & Cao, R. (2022). Applied artificial intelligence in business: Concepts and cases. Springer International Publishing.

Hasan, I., Rizvi, S. (2022). AI-Driven Fraud Detection and Mitigation in e-Commerce Transactions. In: Gupta, D., Polkowski, Z., Khanna, A., Bhattacharyya, S., Castillo, O. (eds) Proceedings of Data Analytics and Management. Lecture Notes on Data Engineering and Communications Technologies, 90, 403–414. Springer, Singapore. https://doi.org/10.1007/978-981-16-6289-8_34