| 数据挖掘在新疆肺结核区域发病风险建模与预测中的应用 |
| |
| 引用本文： | 张燕1,尹哲1,贺湘焱2,古丽娜扎尔·艾克拜尔1,刘亚洁1,曹明芹1. 数据挖掘在新疆肺结核区域发病风险建模与预测中的应用[J]. 现代预防医学, 2020, 0(4): 583-587 |
| |
| 作者姓名： | 张燕1 尹哲1 贺湘焱2 古丽娜扎尔·艾克拜尔1 刘亚洁1 曹明芹1 |
| |
| 作者单位： | 1. 新疆医科大学公共卫生学院，新疆 乌鲁木齐 830011；2. 新疆维吾尔自治区人民医院，新疆 乌鲁木齐 830001 |
| |
| 摘 要： | 目的 采用数据挖掘技术与方法构建生态学因素预测模型，探讨其在新疆涂阳肺结核区域发病风险(SMR)中的应用价值，为新疆结核病的精准防控提供方法学参考。方法 分别采用Lasso和RFE对生态学指标进行筛选，构建Lasso回归和SVR模型，对比与评价建模效果。结果 新疆西南部地区涂阳肺结核SMR较高, Lasso和RFE特征选择结果存在差异，SVR模型预测效果整体上优于OLS和Lasso回归模型，Lasso法结合SVR模型预测效果最优。结论 依据不同地区、不同生态学现状SMR水平的差异，针对性地采取肺结核的预防与控制措施，对肺结核疫情的精准防控具有重要的实践意义。
|
| 关 键 词： | 涂阳肺结核 空间统计学 特征选择Lasso回归 支持向量回归 |
Application of data mining in modeling and prediction of regional risk of tuberculosis in Xinjiang |
| |
| ZHANG Yan,YIN Zhe,HE Xiang-yan,GULINAZHAER Ai-ke-bai-er,LIU Ya-jie,CAO Ming-qin. Application of data mining in modeling and prediction of regional risk of tuberculosis in Xinjiang[J]. Modern Preventive Medicine, 2020, 0(4): 583-587 |
| |
| Authors: | ZHANG Yan YIN Zhe HE Xiang-yan GULINAZHAER Ai-ke-bai-er LIU Ya-jie CAO Ming-qin |
| |
| Affiliation: | School of public health, Xinjiang Medical University, Urumqi, Xinjiang 830011, China |
| |
| Abstract: | Constructing the ecological factor prediction model with data mining technology and methods, the study a imed to explore its application value in the prediction of regional standardized morbidity ratio(SMR) of smear-positive pulmonary tuberculosis in Xinjiang and provide a reference for the precise prevention and control of tuberculosis in Xinjiang. Methods Lasso and RFE were used to select the ecological indicators, and the Lasso regression and SVR model were constructed to compare and evaluate the modeling effects. Results The regional SMR of smear-positive pulmonary tuberculosis in the southwest Xinjiang was higher. The results of Lasso and RFE feature selection were different. The prediction effect of SVR model was better than that of OLS and Lasso regression model. The Lasso method combined with SVR model had the best prediction effect.Conclusion According to the difference of SMR levels in different regions and different ecological status, the prevention and control measures of tuberculosis can be taken in a targeted manner, which has important practical significance for the accurate prevention and control of tuberculosis epidemic. |
| |
| Keywords: | Smear-positive pulmonary tuberculosis Spatial statistics Feature selection Lasso regression Support vector regression |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《现代预防医学》浏览原始摘要信息 |
|
点击此处可从《现代预防医学》下载免费的PDF全文 |
