Pairwise Relative Distance (PRED) is an intuitive and robust metric for assessing vector similarity and class separability

Pairwise Relative Distance (PRED) is an intuitive and robust metric for assessing vector similarity and class separability
复制标题

成对相对距离 (PRED) 是一种直观且稳健的指标,用于评估向量相似性和类可分离性

DOI:
10.1101/2021.08.13.456194
复制
发表时间:
2021
期刊:
--
影响因子:
--
通讯作者:
Mittal A
Mittal A
中科院分区:
--
文献类型:
--
作者:
Mittal A

文献摘要

参考文献

相似文献

科学研究通常需要评估有序值集之间的相似性。每一个集合,包含一个值的每一个维度或类的数据,可以方便地表示为一个向量。向量相似性的常用度量包括基于角度的度量,例如余弦相似性或皮尔逊相关性,其比较值的相对模式,以及基于距离的度量,例如欧几里得距离,其比较值的幅度。在这里,我们评估一个新提出的度量,成对相对距离(PRED),它考虑了相对模式和幅度提供一个单一的向量相似性的措施。PRED本质上揭示了向量是否如此相似,以至于它们在类中的值是可分离的。通过比较PRED在各种应用程序中的其他常见指标,我们表明,PRED提供了一个稳定的机会水平,无论类的数量,是不变的全球平移和缩放操作的数据,具有高动态范围和低变异性,在处理噪声数据,并可以处理多维数据,在包含时间或人口响应的向量的情况下,为每个类。我们还发现,PRED可以适应作为类可分性的可靠度量,即使是缺乏向量结构的数据集,只是包含每个类的多个值。
Scientific studies often require assessment of similarity between ordered sets of values. Each set, containing one value for every dimension or class of data, can be conveniently represented as a vector. The commonly used metrics for vector similarity include angle-based metrics, such as cosine similarity or Pearson correlation, which compare the relative patterns of values, and distance-based metrics, such as the Euclidean distance, which compare the magnitudes of values. Here we evaluate a newly proposed metric, pairwise relative distance (PRED), which considers both relative patterns and magnitudes to provide a single measure of vector similarity. PRED essentially reveals whether the vectors are so similar that their values across the classes are separable. By comparing PRED to other common metrics in a variety of applications, we show that PRED provides a stable chance level irrespective of the number of classes, is invariant to global translation and scaling operations on data, has high dynamic range and low variability in handling noisy data, and can handle multi-dimensional data, as in the case of vectors containing temporal or population responses for each class. We also found that PRED can be adapted to function as a reliable metric of class separability even for datasets that lack the vector structure and simply contain multiple values for each class.
DOI: 10.1073/pnas.1500804112
发表时间: 2015-05-26
影响因子: 11.1
作者:
Buchanan, Sean M.;Kain, Jamey S.;de Bivort, Benjamin L.
通讯作者: de Bivort, Benjamin L.
DOI: 10.1073/pnas.1901623116
发表时间: 2020-09-22
影响因子: 11.1
作者:
Honegger, Kyle S.;Smith, Matthew A-Y;de Bivort, Benjamin L.
通讯作者: de Bivort, Benjamin L.
DOI: 10.7554/elife.66018
发表时间: 2021-05-25
期刊: eLife
影响因子: 7.7
作者:
Schlegel P;Bates AS;Stürner T;Jagannathan SR;Drummond N;Hsu J;Serratosa Capdevila L;Javier A;Marin EC;Barth-Maron A;Tamimi IF;Li F;Rubin GM;Plaza SM;Costa M;Jefferis GSXE
通讯作者: Jefferis GSXE
DOI: --
发表时间: 2018
期刊:
影响因子: --
作者:
金澤建彦;上田貴志;金澤建彦;Kazumichi Shimizu and Mark Stopfer
通讯作者: Kazumichi Shimizu and Mark Stopfer
决定信:高级嗅觉中枢的功能和解剖特异性
DOI: 10.7554/elife.44590.025
发表时间: 2019
期刊: eLife
影响因子: 7.7
作者:
I. G. Kadow
通讯作者: I. G. Kadow