The Research of Clinical Decision Support System Based on Three-Layer Knowledge Base Model.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine

Background:

• Traditional clinical decision support systems often use a two-layer (disease-symptom) knowledge base, which can limit inference accuracy.
• This model's simplicity in inferring diseases solely from symptoms may not fully capture complex medical knowledge.

Purpose of the Study:

• To propose and evaluate a novel three-layer knowledge base model (disease-symptom-property) for enhanced clinical decision support.
• To improve the accuracy and expressiveness of medical knowledge representation in decision support systems.

Main Methods:

• Developed a three-layer knowledge base model incorporating disease, symptom, and property information.
• Utilized a multisymptom naive Bayes algorithm with weighted symptom specificity for probability calculation.
• Implemented an online learning process for parameter optimization of the inference engine.

Main Results:

• The three-layer model demonstrated improved prediction accuracy compared to the conventional two-layer model.
• The system provided effective clinical recommendations, validated by experienced physicians against clinical outcomes.
• Weighted symptom specificity and reduced attribute dependencies enhanced the naive Bayes algorithm's applicability.

Conclusions:

• The proposed three-layer knowledge base model offers a more effective approach to clinical decision support.
• This model enhances prediction accuracy and provides valuable clinical recommendations.
• Further research is needed to address identified limitations and expand system capabilities.