Centers for Biomedical Engineering, University of Science and Technology of China, Hefei, Anhui 230027, China.
Medical Imaging Center, The First Affiliated Hospital of Anhui University of Chinese Medicine, Hefei 230031, China.
J Healthc Eng. 2017;2017:6535286. doi: 10.1155/2017/6535286. Epub 2017 Jul 27.
In many clinical decision support systems, a two-layer knowledge base model (disease-symptom) of rule reasoning is used. This model often does not express knowledge very well since it simply infers disease from the presence of certain symptoms. In this study, we propose a three-layer knowledge base model (disease-symptom-property) to utilize more useful information in inference. The system iteratively calculates the probability of patients who may suffer from diseases based on a multisymptom naive Bayes algorithm, in which the specificity of these disease symptoms is weighted by the estimation of the degree of contribution to diagnose the disease. It significantly reduces the dependencies between attributes to apply the naive Bayes algorithm more properly. Then, the online learning process for parameter optimization of the inference engine was completed. At last, our decision support system utilizing the three-layer model was formally evaluated by two experienced doctors. By comparisons between prediction results and clinical results, our system can provide effective clinical recommendations to doctors. Moreover, we found that the three-layer model can improve the accuracy of predictions compared with the two-layer model. In light of some of the limitations of this study, we also identify and discuss several areas that need continued improvement.
在许多临床决策支持系统中，使用了两层知识库模型（疾病-症状）的规则推理。由于该模型只是根据某些症状推断疾病，因此通常不能很好地表达知识。在本研究中，我们提出了一个三层知识库模型（疾病-症状-特性），以利用推理中更有用的信息。该系统基于多症状朴素贝叶斯算法，迭代计算可能患有疾病的患者的概率，其中通过对诊断疾病的贡献程度的估计对这些疾病症状的特异性进行加权。这显著减少了属性之间的依赖性，从而更适当地应用朴素贝叶斯算法。然后，完成了推理引擎参数优化的在线学习过程。最后，由两位经验丰富的医生对我们利用三层模型的决策支持系统进行了正式评估。通过预测结果与临床结果的比较，我们的系统可以为医生提供有效的临床建议。此外，我们发现与两层模型相比，三层模型可以提高预测的准确性。鉴于本研究的一些局限性，我们还确定并讨论了需要继续改进的几个方面。
