社区搜索的目标是从数据图中得到包含查询顶点的紧密子图,在社会学、生物学等领域有着广泛应用。针对现有基于子图连通性的社区模型的基础连通结构都是完全连通图,无法满足实际应用中用户对社区结构多样性的需求的问题,提出一种基于moti...社区搜索的目标是从数据图中得到包含查询顶点的紧密子图,在社会学、生物学等领域有着广泛应用。针对现有基于子图连通性的社区模型的基础连通结构都是完全连通图,无法满足实际应用中用户对社区结构多样性的需求的问题,提出一种基于motif连通性的社区搜索方法,其中包括基于motif连通性的社区(MCC)模型以及两个相应的社区搜索算法——MPCS(Motif-Processed Community Search)算法和基于MP-index的社区搜索算法。MCC模型可以协助用户自由指定社区的基础连通结构,MPCS算法可以用来解决MCC的搜索问题。此外,提出两个分别针对motif实例搜索过程及所属社区判断过程的剪枝优化技术。最后,设计了MP-index以避免社区搜索过程中的冗余遍历操作。在多个真实数据集上进行实验的结果表明:剪枝优化可以使MPCS算法的耗时减少60%~85%,而基于MP-index的社区搜索算法相较于加入剪枝优化的MPCS算法,效率提升普遍达到了2~3个数量级。可见,所提方法在商品推荐和社交网络等问题上有着实际应用价值。展开更多
In Eastern South America, high altitude grasslands represent a mountain system that has a high number of endemic species. However, studies on the ecology of plant communities in these environments remain scarce. We ai...In Eastern South America, high altitude grasslands represent a mountain system that has a high number of endemic species. However, studies on the ecology of plant communities in these environments remain scarce. We aimed to evaluate the patterns of biodiversity and structure of plant communities from rocky outcrops in high altitude grasslands of three areas at the Caparaó National Park, southeastern Brazil, by sampling 300 randomly distributed plots. Then, we compared the floristic composition, relative abundance, and biological and vegetation spectra among areas. We classified species as endemic and non-endemic and verified the occurrence of endangered species. Species richness was evaluated by rarefaction analysis on the sampling units. The importance value and species abundance distribution(SAD) models were assessed. We also performed an indicator species analysis. We sampled 58 species belonging to 49 genera and 32 families. The number of species decreased with increasing altitude, with significant differences being observed among areas regarding richness, abundance, and cover. Of the total number of species, 10 are endemic to the Caparaó National Park and 17 are listed on the Brazilian Red List of endangered species. The dominant families on all peaks were Asteraceae and Poaceae. The SAD models showed lognormal and geometric distributions, corroborating the fact that 10 species that were common to all three areas were also the most dominant ones in the communities and showed the highest importance values, which ranged between 35% and 60%. Indicator species analysis revealed that 28 species(48.27%) were indicators. Of these, 42.85% had maximum specificity, meaning that they occurred only in one area. Thus, the number of species per life form ratio was similar among areas, yet vegetation spectra differed, especially for hemicryptophytes. The altimetric difference among the areas showed to be a very important driver in the community assembly, influencing the evaluated variables, however, other drivers as soil depth, slope and water could also influence the community structure on a smaller and local spatial scale.展开更多
文摘社区搜索的目标是从数据图中得到包含查询顶点的紧密子图,在社会学、生物学等领域有着广泛应用。针对现有基于子图连通性的社区模型的基础连通结构都是完全连通图,无法满足实际应用中用户对社区结构多样性的需求的问题,提出一种基于motif连通性的社区搜索方法,其中包括基于motif连通性的社区(MCC)模型以及两个相应的社区搜索算法——MPCS(Motif-Processed Community Search)算法和基于MP-index的社区搜索算法。MCC模型可以协助用户自由指定社区的基础连通结构,MPCS算法可以用来解决MCC的搜索问题。此外,提出两个分别针对motif实例搜索过程及所属社区判断过程的剪枝优化技术。最后,设计了MP-index以避免社区搜索过程中的冗余遍历操作。在多个真实数据集上进行实验的结果表明:剪枝优化可以使MPCS算法的耗时减少60%~85%,而基于MP-index的社区搜索算法相较于加入剪枝优化的MPCS算法,效率提升普遍达到了2~3个数量级。可见,所提方法在商品推荐和社交网络等问题上有着实际应用价值。
文摘In Eastern South America, high altitude grasslands represent a mountain system that has a high number of endemic species. However, studies on the ecology of plant communities in these environments remain scarce. We aimed to evaluate the patterns of biodiversity and structure of plant communities from rocky outcrops in high altitude grasslands of three areas at the Caparaó National Park, southeastern Brazil, by sampling 300 randomly distributed plots. Then, we compared the floristic composition, relative abundance, and biological and vegetation spectra among areas. We classified species as endemic and non-endemic and verified the occurrence of endangered species. Species richness was evaluated by rarefaction analysis on the sampling units. The importance value and species abundance distribution(SAD) models were assessed. We also performed an indicator species analysis. We sampled 58 species belonging to 49 genera and 32 families. The number of species decreased with increasing altitude, with significant differences being observed among areas regarding richness, abundance, and cover. Of the total number of species, 10 are endemic to the Caparaó National Park and 17 are listed on the Brazilian Red List of endangered species. The dominant families on all peaks were Asteraceae and Poaceae. The SAD models showed lognormal and geometric distributions, corroborating the fact that 10 species that were common to all three areas were also the most dominant ones in the communities and showed the highest importance values, which ranged between 35% and 60%. Indicator species analysis revealed that 28 species(48.27%) were indicators. Of these, 42.85% had maximum specificity, meaning that they occurred only in one area. Thus, the number of species per life form ratio was similar among areas, yet vegetation spectra differed, especially for hemicryptophytes. The altimetric difference among the areas showed to be a very important driver in the community assembly, influencing the evaluated variables, however, other drivers as soil depth, slope and water could also influence the community structure on a smaller and local spatial scale.
