Background CH<sub>4</sub> concentration and seasonal variations measured at Zhongshan Station (69°22'2''S, 76°21'49''E, 18.5 m) in Antarctica from 2008 through 2013 are pr...Background CH<sub>4</sub> concentration and seasonal variations measured at Zhongshan Station (69°22'2''S, 76°21'49''E, 18.5 m) in Antarctica from 2008 through 2013 are presented and discussed. From 2008-2013 CH<sub>4</sub> was measured in weekly<sub> </sub>flask samples and started on line measurement by Picarro CO<sub>2</sub>/CH<sub>4</sub>/H<sub>2</sub>O analyzer from March, 2010-2013. These CH<sub>4</sub> measurements show the expected growth period of CH<sub>4</sub> concentration during February (Antarctic spring) with a peak in September (fall). Irrespective of wind direction, CH<sub>4</sub> concentrations distribute evenly after the removal of polluted air from station operations, accounting for 1% of the data. The mean daily cycle of CH<sub>4</sub> concentration in all four seasons is small. The monthly mean CH<sub>4</sub> concentration at Zhongshan station is similar to those at other stations in Antarctica showing that CH<sub>4</sub> observed in Antarctica is fully mixed in the atmosphere as it is transported from the northern through the southern hemisphere. The annual CH<sub>4</sub> increase in recent years at Zhongshan station is 4.8 ppb·yr<sup>-1</sup>.展开更多
针对传统的K-medoids聚类算法在聚类时需要随机选择初始类中心且指定聚类数目K,及聚类结果不稳定的问题,提出了一种优化初始类中心的自适应K-medoids算法(adaptive K-medoids algorithm for optimizing initial class centers,CH_KD)....针对传统的K-medoids聚类算法在聚类时需要随机选择初始类中心且指定聚类数目K,及聚类结果不稳定的问题,提出了一种优化初始类中心的自适应K-medoids算法(adaptive K-medoids algorithm for optimizing initial class centers,CH_KD).其思想是定义了特征重要度,以此筛选出每一簇中最优的代表特征,组成特征子集,并重点研究了传统划分算法的自适应优化与改进.首先,利用特征标准差定义特征区分度,选择出区分度强的特征.其次,利用皮尔逊相关系数度量特征簇中每个特征的冗余度,选择出冗余度低的特征.最后,将特征区分度与特征冗余度之积作为特征重要度,以此筛选出每一簇中最优的代表特征,组成特征子集.实验将所提算法与其他聚类算法在14个UCI数据集上进行对比,结果验证了CH_KD算法的有效性与优势.展开更多
文摘Background CH<sub>4</sub> concentration and seasonal variations measured at Zhongshan Station (69°22'2''S, 76°21'49''E, 18.5 m) in Antarctica from 2008 through 2013 are presented and discussed. From 2008-2013 CH<sub>4</sub> was measured in weekly<sub> </sub>flask samples and started on line measurement by Picarro CO<sub>2</sub>/CH<sub>4</sub>/H<sub>2</sub>O analyzer from March, 2010-2013. These CH<sub>4</sub> measurements show the expected growth period of CH<sub>4</sub> concentration during February (Antarctic spring) with a peak in September (fall). Irrespective of wind direction, CH<sub>4</sub> concentrations distribute evenly after the removal of polluted air from station operations, accounting for 1% of the data. The mean daily cycle of CH<sub>4</sub> concentration in all four seasons is small. The monthly mean CH<sub>4</sub> concentration at Zhongshan station is similar to those at other stations in Antarctica showing that CH<sub>4</sub> observed in Antarctica is fully mixed in the atmosphere as it is transported from the northern through the southern hemisphere. The annual CH<sub>4</sub> increase in recent years at Zhongshan station is 4.8 ppb·yr<sup>-1</sup>.
文摘针对传统的K-medoids聚类算法在聚类时需要随机选择初始类中心且指定聚类数目K,及聚类结果不稳定的问题,提出了一种优化初始类中心的自适应K-medoids算法(adaptive K-medoids algorithm for optimizing initial class centers,CH_KD).其思想是定义了特征重要度,以此筛选出每一簇中最优的代表特征,组成特征子集,并重点研究了传统划分算法的自适应优化与改进.首先,利用特征标准差定义特征区分度,选择出区分度强的特征.其次,利用皮尔逊相关系数度量特征簇中每个特征的冗余度,选择出冗余度低的特征.最后,将特征区分度与特征冗余度之积作为特征重要度,以此筛选出每一簇中最优的代表特征,组成特征子集.实验将所提算法与其他聚类算法在14个UCI数据集上进行对比,结果验证了CH_KD算法的有效性与优势.
