G-Mode Classification of Trans Neptunian Objects
G-Mode Classification of Trans Neptunian Objects
复制标题
海王星外天体的 G 模式分类
DOI：
--
复制
发表时间：
2006
期刊：
影响因子：
--
通讯作者：
A. Gavrishin
中科院分区：
文献类型：
--
作者：
M. D. Sanctis;A. Coradini;A. Gavrishin
简介：TNO种群表现出广泛的颜色多样性。现在人们普遍认为，颜色多样性是外太阳系种群的一个真实的特征，而不是观测偏差。颜色范围从蓝色到非常红色，但TNO的颜色分布如何仍在争论中。由于总体的特征是由几个参数，反映不同的现象，重要的是要使用多变量统计，以了解是否存在不同类型的对象。在下文中，我们将首先描述人口的特征，然后描述解决这一问题所需的战略。TNOs数据样本：我们使用81个柯伊伯带天体的数据进行统计分析。我们在具有众所周知的动力学元素的KBO中选择了我们的样本：所有的对象都是在具有计算出的适当元素的对象之间选择的（这意味着两个或更多的反对派）[1]。在我们的统计分析中使用的参数是：颜色指数B-V和V-R，星等H，轨道倾角i，轨道偏心率e和半长轴a。颜色取自不同的发布数据。G-模式：我们使用G-模式多元统计方法来分析这个数据集。Gavrishin和A. Coradini [2].作为一种聚类方法，G模式允许用户将由N个样本描述的统计宇宙（每个样本取决于M个变量）分类为同质分类单元，考虑仪器误差，并且如果存在，则消除冗余变量。该方法的另一个优点是可以执行不同级别的分类。一旦发现第一个准同质群，进一步的分析通常可以将它们细分为同质类。在这一点上，还可以识别类之间的相似性和差异性的水平。结果：在99.5%置信水平的条件下，我们将G-模式方法应用于我们的样本。G-模式分析将我们样本中的81个物体分为五组，动力学参数分离良好，颜色分离较少。五种类型中的四种，1，2，4和5，在动力学参数上分离良好，而类型3显示出广泛的参数范围，并与类型1和2叠加（图1）。类型4和类型5在半长轴方面的结果更极端（类型4和类型5的平均值分别为a = 58.3和a = 89.4）。
Introduction: TNO population show a wide colour diversity. Now is quite widely accepted that the colour diversity is a real characteristics of the outer solar system population and not an observational bias. Colors range from blue to very red but the how is colour distribution of TNOs is under debate. Since the population is characterized by several parameters, reflecting the different phenomena, it is important to use a multivariate statistics in order to understand if different types of objects exist. In what follows we will describe first the characteristic of the population, and then we will describe the strategy needed to tackle this problem. TNOs Data Sample: We make a statistical analysis using the data of 81 Kuiper Belt objects. We chosen our sample among the KBOs with well known dynamical elements: all the objects have been chosen between those with calculated proper elements (that means two or more oppositions) [1] . The parameters used in our statistical analysis are: colour index B-V and V-R, magnitude H, orbital inclination i, orbital eccentricity e, and semimajor axes a. The colors are taken from different published data. G-mode: We used the G-mode multivariate statistical approach to analyse this data set The G-mode method was developed by A. Gavrishin and A. Coradini [2]. As a clustering method, the G-mode allows the user to classify a statistical universe described by N samples, each depending on M variables, into homogeneous taxonomic units, considering the instrumental errors, and, if present, to eliminate redundant variables. Another advantage of the method is that different levels of classification can be performed. Once the first quasi-homogeneous groups are found, further analysis can usually subdivide them in homogeneous classes. At this point, it is also possible to recognize the level of similarity and differences between classes. Results: We applied the G-mode method to our sample the conditions of 99.5% of confidence level. G-mode analysis separates the 81 objects of our sample in five groups, well separated for the dynamical parameters and less separated in colours. Four of the five type, 1, 2, 4, and 5, are well separated in dynamical parameters while type 3 show a wide range of parameters and is superimposed to type 1 and 2 (fig.1). Types 4 and 5 results as the more extreme in terms of semi major axes (average a = 58.3 and a = 89.4 for type 4 and 5 respectively).