基于GIS熵指数与Logistic回归模型的地裂缝灾害评价研究

Evaluation of Ground Fissure Disasters using Index of Entropy and Logistic Regression Models in GIS

选取地层岩性、地质构造、地下水、地貌和人类活动5个区域地裂缝灾害影响因素,采用确定性系数(CF)分析方法,建立了熵指数(IOE)模型和Logistic回归模型,运用GIS技术构建地裂缝灾害评价体系,研究在GIS环境下地裂缝灾害危险等级划分方法,并对苏锡常地区地裂缝灾害进行评价。采用严重、较严重、一般和不严重4种等级划分灾害等级,其中熵指数模型评价结果中各等级占总灾害面积比分别为:10.41%、13.21%、38.11%、38.27%,准确度为92%。Logistic回归模型评价结果为:7.48%、13.33%、34.85%、44.34%,准确度为84%。两种模型评价的灾害结果说明,熵指数模型和Logistic回归模型均可用作地裂缝灾害评价模型,通过算例证明,熵指数模型结果准确可靠。

This paper selects lithology,geological structure,groundwater,geomorphology and human engineering activities as the factors in ground fissure disasters. Using bivariate（ certainty factor） method,the entropy index（ IOE） model and Logistic regression model are established. Then ArcGIS is used to build the evaluation system of ground fissure disasters. With the ground fissure hazard classification,the disaster level figure in GIS environment drawn. Taking Suzhou,Wuxi and Changzhou as the study cases the regional ground fissure assessment is carried out. Ground fissure disasters are divided into 4 grades of serious,more serious,general and not severe and the result of the entropy index of model evaluation shows the 4 grades occupy 10. 41%,13. 21%,38. 11% and 38. 27% of the total disaster area,respectively while the logistic regression model evaluation result shows 7. 48%,13. 33%,34. 85% and 44. 34% respectively. The accuracies of the two models are 92% and84%,respectively. The results of the two models are similar or in line with the actual. So both the entropy index model and the logistic regression model can be used to evaluate ground fissure disasters. But in contrast with actual data,the results of the entropy index model are more accurate.