灾害学2017，Vol.32Issue(1)：17-21,5.DOI:10.3969/j.issn.1000-811X.2017.01.004
基于多分类logistic模型的铁路水害分级警戒概率预报研究
Graded Alerting Probability Forecast of Railway Water Disaster Based on the Multinomial Logistic Regression Model
摘要
Abstract
Based on the railway water disaster data and precipitation data of Yingxia Railway from 2004 to 2014,the temporal and spatial distribution characteristics of the railway water disaster are analyzed,and the graded alerting probability forecast model of railway water disaster by two different terrain are researched.The results show that the annual variation and distribution of flood disaster are different,railway water disaster always occurred in flood season.The test results of probability forecast model by mountain terrain show that the graded alerting accura-cy rates are 78.3%,58.8% and 74.4%.The test results of probability forecast model by relatively flat terrain show that the graded alerting accuracy rates were 85.0%,75.8%,and 87.7%.Alerting effect could be better when combined with the probability forecast model and the precipitation values of 3 factors,the result have an im-portant guidance for the safe operation and efficient operation of the railway department.关键词
鹰厦铁路/铁路水害/分级警戒/降水/多分类logistic回归Key words
Yingxia Railway/railway water disaster/graded alertness/Precipitation Multinomial Logistic Re-gression分类
资源环境引用本文复制引用
吴凡,阙志萍..基于多分类logistic模型的铁路水害分级警戒概率预报研究[J].灾害学,2017,32(1):17-21,5.基金项目
江西省科技厅项目（20142BBG70035）；南昌铁路局科委项目“局管内水害临界雨量及评估模型研究” （）
