Science Enabled by Specimen Data

随机森林(Random forest)模型在2001年发表后得到广泛的关注。由于随机森林可以进行回归和判别等多种统计分析,而且不受正态性、方差齐性和自变量独立性等参数检验的前提条件的制约,其应用日益普遍,有被看作万能模型的趋势。实际上,随机森林是一种特点鲜明的模型,应用局部优化拟合观察值,在分析有偏效应关系的数据时,其结果往往不准确。本文以蝉科(Cicadidea)物种的分布数据为例,比较了随机森林在回归分析时与多元线性回归、广义可加模型和人工神经网络模型的差别,在判别分析时与线性判别分析的差别,强调了随机森林预测时的碎片化特点。结果显示随机森林在处理有多元共线性和交互作用的数据时,以及在判别…

Li, X., Li, B., Wang, G., Zhan, X., & Holyoak, M. (2020). Deeply digging the interaction effect in multiple linear regressions using a fractional-power interaction term. MethodsX, 7, 101067. doi:10.1016/j.mex.2020.101067 https://doi.org/10.1016/j.mex.2020.101067

In multiple regression Y ~ β0 + β1X1 + β2X2 + β3X1 X2 + ɛ., the interaction term is quantified as the product of X1 and X2. We developed fractional-power interaction regression (FPIR), using βX1M X2N as the interaction term. The rationale of FPIR is that the slopes of Y-X1 regression along the X2 gr…

Piel, W. H. (2018). The global latitudinal diversity gradient pattern in spiders. Journal of Biogeography, 45(8), 1896–1904. doi:10.1111/jbi.13387 https://doi.org/10.1111/jbi.13387

Aim: The aim of this study was to test the hypothesis that the global latitudinal diversity gradient pattern in spiders is pear‐shaped, with maximum species diversity shifted south of the Equator, rather than egg‐shaped, centred on the equator, this study infers the gradient using two large datasets…