Ranking network of a captive rhesus macaque society: a sophisticated corporative kingdom.

Ranking network of a captive rhesus macaque society: a sophisticated corporative kingdom.
复制标题

俘虏恒河猕猴社会的排名网络:一个复杂的企业王国。

DOI:
10.1371/journal.pone.0017817
复制
发表时间:
2011-03-15
期刊:
影响因子:
3.7
通讯作者:
McCowan B
McCowan B
中科院分区:
综合性期刊3区
文献类型:
--
作者:
Fushing H;McAssey MP;Beisner B;McCowan B

文献摘要

参考文献

被引文献

相似文献

我们开发了一个三步计算的方法来探索一个社会的圈养恒河猴的层次排名网络。计算的网络是足够的信息来解决这个问题:是恒河猴社会的排名网络更像一个王国或公司?我们的计算是基于一个三步的方法。这些步骤的目的是处理的巨大挑战所产生的传递性的优势,作为一个必要的约束所有个体猕猴之间的排名关系,以及非常高的采样异质性的行为冲突数据。第一步同时推断所有网络成员之间的排名潜力,这需要适应行为数据中固有的异构测量误差。我们的第二步通过最小化排名潜力中的网络范围误差来估计所有个体的社会排名。第三步提供了一种方法来计算社会排名中选定的经验特征的置信界限。我们采用这种方法的两套冲突的数据有关的两个圈养社会的成年恒河猴。由此产生的排名网络,每个社会被发现是一个复杂的混合物,既一个王国和公司。此外,为了验证的目的,我们重新分析了20个长角羊的冲突数据,并证明了我们的三步方法是能够正确计算的排名网络,消除所有的排名错误。
We develop a three-step computing approach to explore a hierarchical ranking network for a society of captive rhesus macaques. The computed network is sufficiently informative to address the question: Is the ranking network for a rhesus macaque society more like a kingdom or a corporation? Our computations are based on a three-step approach. These steps are devised to deal with the tremendous challenges stemming from the transitivity of dominance as a necessary constraint on the ranking relations among all individual macaques, and the very high sampling heterogeneity in the behavioral conflict data. The first step simultaneously infers the ranking potentials among all network members, which requires accommodation of heterogeneous measurement error inherent in behavioral data. Our second step estimates the social rank for all individuals by minimizing the network-wide errors in the ranking potentials. The third step provides a way to compute confidence bounds for selected empirical features in the social ranking. We apply this approach to two sets of conflict data pertaining to two captive societies of adult rhesus macaques. The resultant ranking network for each society is found to be a sophisticated mixture of both a kingdom and a corporation. Also, for validation purposes, we reanalyze conflict data from twenty longhorn sheep and demonstrate that our three-step approach is capable of correctly computing a ranking network by eliminating all ranking error.
DOI: 10.1007/s002650000222
发表时间: 2000-11-01
影响因子: 2.3
作者:
Goessmann, C;Hemelrijk, C;Huber, R
通讯作者: Huber, R
DOI: 10.1163/156853902760102645
发表时间: 2002-02-01
期刊: BEHAVIOUR
影响因子: 1.3
作者:
Isbell, LA;Young, TP
通讯作者: Young, TP
DOI: 10.1163/156853993x00290
发表时间: 1993-06-01
期刊: BEHAVIOUR
影响因子: 1.3
作者:
DREWS, C
通讯作者: DREWS, C
DOI: 10.1016/j.anbehav.2004.08.011
发表时间: 2005-05-01
期刊: ANIMAL BEHAVIOUR
影响因子: 2.5
作者:
Adams, ES
通讯作者: Adams, ES
DOI: 10.1126/science.220.4598.671
发表时间: 1983-01-01
期刊: SCIENCE
影响因子: 56.9
作者:
KIRKPATRICK, S;GELATT, CD;VECCHI, MP
通讯作者: VECCHI, MP