摘要
目的：为了提高计算机病例分型的准确率。方法：模拟专家思维方式，应用多元统计分析方法，构建分型判别方程并增设“校正线”。结果：计算机病例分型与专家分型吻合率达到９７．８％。结论：经４６０万份大样本病例测试和一百多所医院推广应用取得良好效果。
Objective: To enhance the accuracy of case classification by computer. Methods: By simulating expert thinking and applying multi-dimensional statistics, a formula of classification differentiation and a revision curve are set up. Results: The agreement between computer classification and expert classification is 97. 8%. Conclusion: Favorable effects are achieved by applying 4. 6 million samples in over 100 hospitals.
出处
《解放军医院管理杂志》
2000年第2期116-119,共4页
Hospital Administration Journal of Chinese People's Liberation Army
