论文部分内容阅读
本文基于条件独立的贝叶斯公式,计算机辅助预测上消化道再出血。同时对症状变量间的交互作用给予适当的考虑,以探讨不同的方法改善贝叶斯公式在计算机辅助诊断中的判别效果。结果表明,结合聚类分析的条件独立的贝叶斯公式和最佳依赖树的贝叶斯公式,其判别符合率分别比条件独立的贝叶斯公式提高5％(P>0.2)和10％(P<0.05)左右。证明利用最佳依赖树的贝叶斯公式对改善贝叶斯公式的判别效果具有一定的成效。
This article is based on the conditional independent Bayesian formula and predicts the re-bleeding of the upper gastrointestinal tract by computer. At the same time, appropriate considerations are given to the interaction between symptom variables to explore the different methods to improve the discriminant effect of Bayesian formula in computer-aided diagnosis. The results show that the combination of the condition-independent Bayesian formula and the Bayesian formula of the best-dependent tree can increase the discriminant coincidence rate by 5% (P>0.2) and 10%, respectively, compared with the conditional independent Bayesian formula. (P<0.05) or so. It is proved that the Bayesian formula using the optimal dependency tree has certain effects on the improvement of the Bayesian formula.