Although increased risk for adverse birth outcomes has been associated with neighborhood socioeconomic disadvantage, most studies have used cross-sectional measures to account for neighborhood context. Consequently, d...Although increased risk for adverse birth outcomes has been associated with neighborhood socioeconomic disadvantage, most studies have used cross-sectional measures to account for neighborhood context. Consequently, dynamic neighborhood processes that may influence adverse birth outcomes are not fully understood. In this study, a longitudinal measure of socioeconomic change was used to explore variation in low birth weight (LBW) rates between 1990 and 2006 in Chicago neighborhoods. A crosss-ectional measure of neighborhood socioeconomic characteristics was then used to compare the LBW rates across Chicago neighborhoods during the same time frame to determine whether the cross-sectional measure would capture the same nuances in LBW variation as the longitudinal measure. Consistent with previous studies, both measures identified higher low birth weight rates in neighborhoods entrenched in poverty during the study period. However, the longitudinal measure showed that mothers residing in low income neighborhoods with high concentrations of immigrants had LBW rates that were lower than mothers residing in high income neighborhoods. Our results suggest that while cross-sectional measures of neighborhood socioeconomic context may capture global variations in low birth weight rates, longitudinal measures may illuminate subtleties between neighborhoods that might provide an opportunity for targeted policies to reduce adverse maternal and child health outcomes.展开更多
近年来,混合型数据的聚类问题受到广泛关注。作为处理混合型数据的一种有效方法,K-prototype聚类算法在初始化聚类中心时通常采用随机选取的策略,然而这种策略在很多实际应用中难以保证聚类结果的质量。针对上述问题,采用基于离群点检...近年来,混合型数据的聚类问题受到广泛关注。作为处理混合型数据的一种有效方法,K-prototype聚类算法在初始化聚类中心时通常采用随机选取的策略,然而这种策略在很多实际应用中难以保证聚类结果的质量。针对上述问题,采用基于离群点检测的策略来为K-prototype算法选择初始中心,并提出一种新的混合型数据聚类初始化算法(initialization of K-prototype clustering based on outlier detection and density,IKP-ODD)。给定一个候选对象,IKP-ODD通过计算其距离离群因子、加权密度以及与已有初始中心之间的加权距离来判断候选对象是否是一个初始中心。IKP-ODD通过采用距离离群因子和加权密度,防止选择离群点作为初始中心。在计算对象的加权密度以及对象之间的加权距离时,采用邻域粗糙集中的粒度邻域熵来计算每一个属性的重要性,并根据属性重要性的大小为不同属性赋予不同的权重,有效地反映不同属性之间的差异性。在多个UCI数据集上的实验表明,相对于现有的初始化方法,IKP-ODD能够更好地解决K-prototype聚类的初始化问题。展开更多
文摘Although increased risk for adverse birth outcomes has been associated with neighborhood socioeconomic disadvantage, most studies have used cross-sectional measures to account for neighborhood context. Consequently, dynamic neighborhood processes that may influence adverse birth outcomes are not fully understood. In this study, a longitudinal measure of socioeconomic change was used to explore variation in low birth weight (LBW) rates between 1990 and 2006 in Chicago neighborhoods. A crosss-ectional measure of neighborhood socioeconomic characteristics was then used to compare the LBW rates across Chicago neighborhoods during the same time frame to determine whether the cross-sectional measure would capture the same nuances in LBW variation as the longitudinal measure. Consistent with previous studies, both measures identified higher low birth weight rates in neighborhoods entrenched in poverty during the study period. However, the longitudinal measure showed that mothers residing in low income neighborhoods with high concentrations of immigrants had LBW rates that were lower than mothers residing in high income neighborhoods. Our results suggest that while cross-sectional measures of neighborhood socioeconomic context may capture global variations in low birth weight rates, longitudinal measures may illuminate subtleties between neighborhoods that might provide an opportunity for targeted policies to reduce adverse maternal and child health outcomes.
文摘近年来,混合型数据的聚类问题受到广泛关注。作为处理混合型数据的一种有效方法,K-prototype聚类算法在初始化聚类中心时通常采用随机选取的策略,然而这种策略在很多实际应用中难以保证聚类结果的质量。针对上述问题,采用基于离群点检测的策略来为K-prototype算法选择初始中心,并提出一种新的混合型数据聚类初始化算法(initialization of K-prototype clustering based on outlier detection and density,IKP-ODD)。给定一个候选对象,IKP-ODD通过计算其距离离群因子、加权密度以及与已有初始中心之间的加权距离来判断候选对象是否是一个初始中心。IKP-ODD通过采用距离离群因子和加权密度,防止选择离群点作为初始中心。在计算对象的加权密度以及对象之间的加权距离时,采用邻域粗糙集中的粒度邻域熵来计算每一个属性的重要性,并根据属性重要性的大小为不同属性赋予不同的权重,有效地反映不同属性之间的差异性。在多个UCI数据集上的实验表明,相对于现有的初始化方法,IKP-ODD能够更好地解决K-prototype聚类的初始化问题。
