论文部分内容阅读
目的利用分类技术探讨全身炎性反应综合征不同严重程度的诊断规则。方法对64例急性胰腺炎患者的体温、心率、呼吸频率、白细胞数等指标进行分析,利用数据挖掘软件-ROSETTA数据分析工具包中的基于Johnson′s Algorithm的粗糙集约简算法以及统计软件SPSS的基于CHAID的决策树技术进行分类处理。结果基于Johnson′s Algorithm的粗糙集方法得到4条有诊断价值的典型规则,所得规则以体温、心率、呼吸频率和白细胞数4个属性为判断依据;基于CHAID的决策树技术也产生4条有诊断价值的典型规则,所得规则以体温、心率、呼吸频率和白细胞数4个属性中的2个、3个或4个属性为判断依据。全身炎性反应综合征的严重程度可以通过上述4个属性的取值来判断。结论分类技术在判断全身炎性反应综合征的严重程度中有应用价值。
Objective To explore the diagnostic rules of different severity of systemic inflammatory response syndrome by using classification technology. Methods Sixty-four patients with acute pancreatitis were analyzed for body temperature, heart rate, respiration rate, white blood cell count and other indicators. Using the data mining software-ROSETTA data analysis toolkit, the algorithm based on Johnson’s Algorithm and the statistical software SPSS CHAID-based decision tree technology for classification. Results Based on Johnson’s Algorithm, four typical rules with diagnostic value were obtained, and the rules were based on the four attributes of body temperature, heart rate, respiratory rate and white blood cell count. The decision tree based on CHAID also produced four Diagnostic value of the typical rules, the rules derived from the body temperature, heart rate, respiratory rate and white blood cell count of four attributes of 2, 3 or 4 attributes to determine the basis. The severity of systemic inflammatory response syndrome can be judged by the value of the above four attributes. Conclusion Classification technique has the value of judging the severity of systemic inflammatory response syndrome.