Applying Naive Bayesian Networks to Disease Prediction: a Systematic Review

Area of Science:

• Medical Informatics
• Machine Learning
• Biostatistics

Background:

• Naive Bayesian networks (NBNs) are recognized for their simplicity and effectiveness in predictive modeling.
• Their application in disease prediction warrants a comprehensive review of existing evidence.

Purpose of the Study:

• To systematically review the application of NBNs in disease prediction.
• To evaluate the performance of NBNs against other algorithms in medical prognostics.

Main Methods:

• A systematic literature search was conducted on PubMed for articles published between 2005 and 2015.
• Inclusion criteria focused on NBNs and their impact on disease prediction, resulting in the selection of 23 studies after rigorous screening.
• Data extraction included performance metrics such as accuracy, sensitivity, specificity, and Area Under the ROC Curve (AUC).

Main Results:

• The review analyzed 23 studies involving 53,725 patients.
• NBNs consistently showed the best performance across most diseases when compared to alternative predictive algorithms.
• Accuracy was a key metric demonstrating the superiority of NBNs in disease prediction.

Conclusions:

• NBNs are proposed as an efficient method for constructing disease prediction models.
• The findings support NBNs as a leading algorithm for disease prediction, outperforming others in most evaluated cases.
• This systematic review underscores the value of NBNs in enhancing diagnostic accuracy and patient outcomes.