目前对Web服务QoS(Quality of Service)的预测研究,通常预测QoS的静态值,很少预测QoS值的置信区间.本文借助非参数统计学的Bootstrap技术,提出估计Web服务QoS值置信区间的方法;然后利用与当前Web用户相似的其他Web用户调用待预测Web服务...目前对Web服务QoS(Quality of Service)的预测研究,通常预测QoS的静态值,很少预测QoS值的置信区间.本文借助非参数统计学的Bootstrap技术,提出估计Web服务QoS值置信区间的方法;然后利用与当前Web用户相似的其他Web用户调用待预测Web服务的QoS历史数据,预测当前Web用户调用待预测Web服务的QoS值的置信区间.本文估计了WSDream数据集1中每个用户调用每个Web服务的QoS值的置信区间,实验发现这些置信区间的上下限近似服从重尾分布.通过随机选择WSDream数据集1中60%到90%的用户和Web服务作为训练集,预测另外10%到40%的用户和Web服务的QoS值,实验结果表明预测的QoS置信区间与估计的QoS置信区间的平均覆盖率超过70%,最高达76%.在服务选择或服务推荐时给用户提供一个估计的或预测的QoS置信区间,可以更好地满足用户的个性化需求.展开更多
Environmental and Geo-spatial factors have long been considered as crucial determinants of species composition and distributions. However,quantifying the relative contributions of these factors for the alpine ecosyste...Environmental and Geo-spatial factors have long been considered as crucial determinants of species composition and distributions. However,quantifying the relative contributions of these factors for the alpine ecosystems is lacking. The Tibetan Plateau has a unique ecological environment and vegetation types. Our objectives are to quantify the spatial distributions of plant communities on the Northern Tibetan Alpine grasslands and to explore the relationships between vegetation composition,Geo-spatial factors and environmental factors. We established 63 field plots along a 1200-km gradient on the Northern Tibetan Plateau Alpine Grassland and employed the two-way indicator species analysis（TWINSPAN） and the detrended canonical correspondence analysis（DCCA）. Fourteen communities of alpine grassland were identifiable along the transect and consisted of three vegetation types： Alpine meadow,Alpine steppe,and desert steppe. Vegetation composition and spatial distribution appeared to be largely determined by mean annual precipitation and less influenced by temperature. A large fraction（73.5%） of the variation in vegetation distribution was explained by environmental variables along this transect,somewhat less by Geo-spatial factors（56.3%）. The environmental and Geo-spatial factors explained 29.6% and 12.3% of the total variation,respectively,while their interaction explained 43.9%. Our findings provide strong empirical evidence for explaining biological and environmental synergetic relationships in Northern Tibet.展开更多
文摘目前对Web服务QoS(Quality of Service)的预测研究,通常预测QoS的静态值,很少预测QoS值的置信区间.本文借助非参数统计学的Bootstrap技术,提出估计Web服务QoS值置信区间的方法;然后利用与当前Web用户相似的其他Web用户调用待预测Web服务的QoS历史数据,预测当前Web用户调用待预测Web服务的QoS值的置信区间.本文估计了WSDream数据集1中每个用户调用每个Web服务的QoS值的置信区间,实验发现这些置信区间的上下限近似服从重尾分布.通过随机选择WSDream数据集1中60%到90%的用户和Web服务作为训练集,预测另外10%到40%的用户和Web服务的QoS值,实验结果表明预测的QoS置信区间与估计的QoS置信区间的平均覆盖率超过70%,最高达76%.在服务选择或服务推荐时给用户提供一个估计的或预测的QoS置信区间,可以更好地满足用户的个性化需求.
基金National Key Technology Research and Development Program of China(2016YFC0501802,2017YFA0604802)National Natural Science Foundation of China(41571195,41501103)Youth Innovation Team Project of Key Laboratory of Ecosystem Network Observation and Modeling(LENOM2016Q0002)
文摘Environmental and Geo-spatial factors have long been considered as crucial determinants of species composition and distributions. However,quantifying the relative contributions of these factors for the alpine ecosystems is lacking. The Tibetan Plateau has a unique ecological environment and vegetation types. Our objectives are to quantify the spatial distributions of plant communities on the Northern Tibetan Alpine grasslands and to explore the relationships between vegetation composition,Geo-spatial factors and environmental factors. We established 63 field plots along a 1200-km gradient on the Northern Tibetan Plateau Alpine Grassland and employed the two-way indicator species analysis（TWINSPAN） and the detrended canonical correspondence analysis（DCCA）. Fourteen communities of alpine grassland were identifiable along the transect and consisted of three vegetation types： Alpine meadow,Alpine steppe,and desert steppe. Vegetation composition and spatial distribution appeared to be largely determined by mean annual precipitation and less influenced by temperature. A large fraction（73.5%） of the variation in vegetation distribution was explained by environmental variables along this transect,somewhat less by Geo-spatial factors（56.3%）. The environmental and Geo-spatial factors explained 29.6% and 12.3% of the total variation,respectively,while their interaction explained 43.9%. Our findings provide strong empirical evidence for explaining biological and environmental synergetic relationships in Northern Tibet.
