The digestive tract plays an important role in digestion and the acquisition of food energy. Understanding the impact of abiotic environments on digestive tract morphology is especially important for evolution of dige...The digestive tract plays an important role in digestion and the acquisition of food energy. Understanding the impact of abiotic environments on digestive tract morphology is especially important for evolution of digestive tract across different environments. Here, we investigated altitudinal variation in digestive tract length in the Yunnan Pond Frog （Pelophylaxpleuraden） across five populations ranging from 1413 m to 1935 m a.s.1, in Ningnan County, Sichuan province in western China. Frogs were collected during the breeding season, from 1-5 June 2012. Our results revealed that females had longer digestive tract and relative digestive tract （i.e. digestive tract length / body size） lengths in com- parison to males, on average, but the differences between them decreased with increasing altitude. Digestive tract and relative digestive tract lengths increased with increasing altitude suggesting that a higher proportion of indigestible ma- terials may be consumed at high-altitude sites and result in a relative increase in digestive tract dimensions.展开更多
我国西部山区滑坡灾害频发,精确评估滑坡易发性对地质灾害防治至关重要。结合统计方法与机器学习模型的集成模型已广泛的应用于滑坡易发性评价,然而对其结果的进一步优化仍值得考虑。本文提出一种耦合统计方法、机器学习模型以及聚类算...我国西部山区滑坡灾害频发,精确评估滑坡易发性对地质灾害防治至关重要。结合统计方法与机器学习模型的集成模型已广泛的应用于滑坡易发性评价,然而对其结果的进一步优化仍值得考虑。本文提出一种耦合统计方法、机器学习模型以及聚类算法的综合评价方法,以宁南县为例,研究其对滑坡易发性评价精度的提升效应。该方法首先将信息量法(Information Value, IV)、确定系数法(Certainty Factor, CF)和频率比法(Frequency Ratio, FR)分别与随机森林模型(Random Forest, RF)结合,得到三种集成模型(IV-RF、CF-RF、FR-RF)。此后,引入ISO聚类算法对三种集成模型的结果进行分级,得到三种耦合模型(IV-RF-ISO、CF-RF-ISO、FR-RF-ISO)。AUC值(Area Under the Curve)、准确率、F1分数和种子单元面积指数(Seed Cell Area Indexes,SCAI)被用于评估模型的精度。结果显示,集成模型性能均优于单一模型,其准确率和F1分数均大于0.85,AUC值均大于0.9。其中FR-RF模型表现最优,准确率(0.911)、F1分数(0.912)和AUC值(0.965)较FR模型分别提升了0.095、0.096和0.074。与自然断点法和Kmeans聚类法相比,引入ISO算法的耦合模型FR-RF-ISO分级效果最优,其高低易发区SCAI值的差异更为显著。本研究成果表明,耦合统计方法、机器学习与聚类算法的综合评价方法具有较高精度,为提高滑坡易发性评价精度提供思路。展开更多
基金Financial support was provided by the National Natural Sciences Foundation of China (31101633)
文摘The digestive tract plays an important role in digestion and the acquisition of food energy. Understanding the impact of abiotic environments on digestive tract morphology is especially important for evolution of digestive tract across different environments. Here, we investigated altitudinal variation in digestive tract length in the Yunnan Pond Frog （Pelophylaxpleuraden） across five populations ranging from 1413 m to 1935 m a.s.1, in Ningnan County, Sichuan province in western China. Frogs were collected during the breeding season, from 1-5 June 2012. Our results revealed that females had longer digestive tract and relative digestive tract （i.e. digestive tract length / body size） lengths in com- parison to males, on average, but the differences between them decreased with increasing altitude. Digestive tract and relative digestive tract lengths increased with increasing altitude suggesting that a higher proportion of indigestible ma- terials may be consumed at high-altitude sites and result in a relative increase in digestive tract dimensions.
文摘我国西部山区滑坡灾害频发,精确评估滑坡易发性对地质灾害防治至关重要。结合统计方法与机器学习模型的集成模型已广泛的应用于滑坡易发性评价,然而对其结果的进一步优化仍值得考虑。本文提出一种耦合统计方法、机器学习模型以及聚类算法的综合评价方法,以宁南县为例,研究其对滑坡易发性评价精度的提升效应。该方法首先将信息量法(Information Value, IV)、确定系数法(Certainty Factor, CF)和频率比法(Frequency Ratio, FR)分别与随机森林模型(Random Forest, RF)结合,得到三种集成模型(IV-RF、CF-RF、FR-RF)。此后,引入ISO聚类算法对三种集成模型的结果进行分级,得到三种耦合模型(IV-RF-ISO、CF-RF-ISO、FR-RF-ISO)。AUC值(Area Under the Curve)、准确率、F1分数和种子单元面积指数(Seed Cell Area Indexes,SCAI)被用于评估模型的精度。结果显示,集成模型性能均优于单一模型,其准确率和F1分数均大于0.85,AUC值均大于0.9。其中FR-RF模型表现最优,准确率(0.911)、F1分数(0.912)和AUC值(0.965)较FR模型分别提升了0.095、0.096和0.074。与自然断点法和Kmeans聚类法相比,引入ISO算法的耦合模型FR-RF-ISO分级效果最优,其高低易发区SCAI值的差异更为显著。本研究成果表明,耦合统计方法、机器学习与聚类算法的综合评价方法具有较高精度,为提高滑坡易发性评价精度提供思路。
