中国实用神经疾病杂志2025，Vol.28Issue(7)：850-854,5.DOI:10.12083/SYSJ.240534
常规临床资料挖掘对卒中后癫痫发作的预测价值
Prediction value of routine clinical data mining for epileptic after stroke
摘要
Abstract
Objective To investigate the predictive value of routine clinical data mining for post-stroke epilepsy.Methods A total of 100 stroke patients admitted to Handan Central Hospital from January 2020 to December 2022 were selected as the objects.Among them,there were 32 patients with epilepsy after stroke in the study group and 68 patients without epilepsy after stroke in the control group.The clinical data of the two groups were collected and compared,and decision tree C5.0 model and artificial neural network(ANN)model were established,and the effectiveness of the two models in predicting post-stroke seizures was compared.Results There were differences in gender,lesion site,NIHSS score,low density lipoprotein(LDL),C-reactive protein(CRP),D-dimer(D-D),neutrophil to lymphocyte ratio(NLR),and platelet to lymphocyte ratio(PLR)between the study group and the control group(P<0.05).In the decision tree C5.0 model,NIHSS score,LDL,and lesion site were the three factors that had the greatest impact on the model.In the ANN model,NIHSS score,lesion location,and PLR were the three factors that had the greatest impact on the model.The accuracy,positive predictive value,negative predictive value,sensitivity,specificity and AUC of decision tree C5.0 model was 88.00%,84.38%,89.71%,78.64%,77.29%and 0.809,respectively.The accuracy,positive predictive value,negative predictive value,sensitivity,specificity and AUC of ANN model was 95.00%,93.75%,95.59%,88.64%,89.52%,and 0.857,respectively.Conclusion In the prediction model of epilepsy after stroke based on routine clinical data mining,the overall performance of ANN model is better than decision tree C5.0 model,which has potential clinical application value.关键词
脑卒中/癫痫/决策树C5.0/人工神经网络/数据挖掘Key words
Stroke/Epilepsy/Decision tree C5.0/Artificial neural network/Data mining分类
医药卫生引用本文复制引用
宋晓园,晁艳艳..常规临床资料挖掘对卒中后癫痫发作的预测价值[J].中国实用神经疾病杂志,2025,28(7):850-854,5.基金项目
邯郸市科学技术研究与发展计划项目(编号:19422083009-3) （编号:19422083009-3）
