R语言Psych包pairs.panels加权相关矩阵绘图报错求助

Hey there! I see exactly what's causing that error when you try to add weights to your pairs.panels plot. Let's break it down and fix it quickly.

The Root Cause #

The wt parameter in pairs.panels expects a matrix or data frame (a 2-dimensional structure), but you're passing a 1-dimensional vector (iris[,4]). When the function tries to index this vector like a data frame/matrix (wt[, c(1:2)]), it throws the "incorrect number of dimensions" error because vectors don't have columns to index.

The Fixes #

You just need to convert your weight vector into a 2-dimensional structure. Here are two simple ways to do this:

Option 1: Convert to a Data Frame

Wrap your weight vector in data.frame():

require(psych) 
# Weighted correlations (fixed with data frame weights)
pairs.panels(iris[ ,1:3], gap = 0, pch = 21, ellipses = TRUE, 
             show.points = TRUE, smoother = TRUE, rug = FALSE, 
             main = '~ Weighted Correlations', cor = TRUE, 
             wt = data.frame(iris[ ,4]) )

Option 2: Convert to a Matrix

Use as.matrix() to turn the vector into a single-column matrix:

# Weighted correlations (fixed with matrix weights)
pairs.panels(iris[ ,1:3], gap = 0, pch = 21, ellipses = TRUE, 
             show.points = TRUE, smoother = TRUE, rug = FALSE, 
             main = '~ Weighted Correlations', cor = TRUE, 
             wt = as.matrix(iris[ ,4]) )

Quick Check #

If you want to confirm the difference, run these lines in your console:

# 1-dimensional vector (causes error)
str(iris[,4])
# 2-dimensional data frame (works)
str(data.frame(iris[,4]))
# 2-dimensional matrix (works)
str(as.matrix(iris[,4]))

Either fix should make your weighted correlation plot run without errors!

内容的提问来源于stack exchange，提问作者Markm0705