arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2110.07967stat.ME
收藏

用于组合数据空间分析的一类新的$α$ -变换

A new class of $α$-transformations for the spatial analysis of Compositional Data

Lucia Clarotto, Denis Allard, Alessandra Menafoglio

摘要:

地理参考成分数据在许多科学领域和空间统计学中非常重要。这项工作解决的问题,提出模型和方法来分析和预测,通过克里金法,这种类型的数据。为此,提出了一类新的变换,称为等距$α$ -变换($α$ -IT),它包含了传统的等距对数比(ILR)变换。它示出的ILR是$α$ -IT的$α$趋于0的极限情况,并且$α=1$对应于数据的线性变换。与ILR不同,当$α>0$时,所提出的转换接受组合中的0。建立了参数$α$的极大似然估计。使用克里金法对$α$ -IT转换数据进行预测,在合成空间成分数据上进行验证,使用由$α$ -IT诱导的几何形状或单纯形中计算的预测分数。土地覆盖数据的应用表明,相对优势的各种方法w.r.t.预测目标取决于组合物是否包含任何零分量。当所有分量都为正时,极限情况(ILR或线性变换)对于所考虑的度量都不是最优的。一个中间几何,对应于$α$ -IT与最大似然估计,更好地描述了数据集的地质统计设置。当具有0的组合物的量不可忽略时,随着$α$减小,变换的一些副作用被放大,从而导致$α$ -IT几何内和单纯形中的度量的克里金性能差。

英文摘要:

Georeferenced compositional data are prominent in many scientific fields and in spatial statistics. This work addresses the problem of proposing models and methods to analyze and predict, through kriging, this type of data. To this purpose, a novel class of transformations, named the Isometric $α$-transformation ($α$-IT), is proposed, which encompasses the traditional Isometric Log-Ratio (ILR) transformation. It is shown that the ILR is the limit case of the $α$-IT as $α$ tends to 0 and that $α=1$ corresponds to a linear transformation of the data. Unlike the ILR, the proposed transformation accepts 0s in the compositions when $α>0$. Maximum likelihood estimation of the parameter $α$ is established. Prediction using kriging on $α$-IT transformed data is validated on synthetic spatial compositional data, using prediction scores computed either in the geometry induced by the $α$-IT, or in the simplex. Application to land cover data shows that the relative superiority of the various approaches w.r.t. a prediction objective depends on whether the compositions contained any zero component. When all components are positive, the limit cases (ILR or linear transformations) are optimal for none of the considered metrics. An intermediate geometry, corresponding to the $α$-IT with maximum likelihood estimate, better describes the dataset in a geostatistical setting. When the amount of compositions with 0s is not negligible, some side-effects of the transformation gets amplified as $α$ decreases, entailing poor kriging performances both within the $α$-IT geometry and for metrics in the simplex.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑