绘制不同形状的子集答案

【问题标题】：Plotting subsets with different shapes绘制不同形状的子集
【发布时间】：2018-02-14 14:58:09
【问题描述】：

快速问题：我有一个数据集，我将其随机分为训练子集和测试子集。然后我做了一些统计分析，想将结果一起绘制在同一个图中，但对两个不同的子集使用不同的形状。

我是 ggplot 的新手，所以我的问题是必须在一开始就为 ggplot 提供完整的数据集。由于我已通过随机索引将数据分成两组，因此我找不到 aes() 的正确选择属性

data=read.csv("...",sep=" ")
data$class = as.factor(data$class - 1)
colnames(data)=c("y","x1","x2")
n = dim(data)[1]
order = sample(n)
test = data[order[1:(n/2)],]
train = data[order[(n/2):n,]
#...
ggplot(train) + geom_point( aes(x =x1, y = x2, color = y)) 
# this should be done for the whole dataset, kinda like this
# ggplot(data) + geom_point(aes(x=x1, y=x2, color=y, shape=(index is in test and not train)))
# which is obviously not valid

感谢您的宝贵时间，尼克拉斯

【问题讨论】：

我们可以看一些示例数据吗？这应该很简单，但希望查看一些数据以确保我们做对了。如果可以，请将其插入 R：dput(head(df,10))。这将为我们提供一个可重复使用的示例。您可能希望对测试和训练 dfs 都这样做。
其实是thisstructure(list(y = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("0", "1"), class = "factor"), x1 = c(15.6, 11.2, 18.6, 16.8, 21, 15.2, 14.6, 17.6, 14, 16), x2 = c(5.64, 4.38, 5.68, 7.8, 4.32, 6.75, 5.25, 5.05, 5.2, 7.22)), .Names = c("y", "x1", "x2"), row.names = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10"), class = "data.frame")的子集
形状变量在哪里？形状变量只是训练集中的 1/0 吗？
没有shape变量，应该是ggplot的参数。形状应通过属于测试集或训练集来确定

标签： r ggplot2

【解决方案1】：

如果您有 2 个数据集，如示例所示，这是另一种方法

library(ggplot2)
data = structure(list(y = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("0", "1"), class = "factor"), x1 = c(15.6, 11.2, 18.6, 16.8, 21, 15.2, 14.6, 17.6, 14, 16), x2 = c(5.64, 4.38, 5.68, 7.8, 4.32, 6.75, 5.25, 5.05, 5.2, 7.22)), .Names = c("y", "x1", "x2"), row.names = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10"), class = "data.frame")
data = data.frame(data)
n = dim(data)[1]
order = sample(n)
test = data[order[1:(n/2)],]
train = data[order[(n/2):n],]

ggplot(test)+geom_point(aes(y = x1,x=x2))+
  geom_point(data = train, aes(y = x1,x = x2), pch = 10, col = "dargreen")

【讨论】：

【解决方案2】：

如果我们假设数据集具有相同的列：

# Loading data    
data <- structure(list(y = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), .Label = c("0", "1"), class = "factor"), x1 = c(15.6, 11.2, 18.6, 16.8, 21, 15.2, 14.6, 17.6, 14, 16), x2 = c(5.64, 4.38, 5.68, 7.8, 4.32, 6.75, 5.25, 5.05, 5.2, 7.22)), .Names = c("y", "x1", "x2"), row.names = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10"), class = "data.frame")
n = dim(data)[1]
order = sample(n)
test = data[order[1:(n/2)],]
train = data[order[(n/2):n,]

test$identifier <- as.factor(1) # mark the test with 1
train$identifier <- as.factor(0) # mark the train with 0

df_out <- rbind(test,train) # combine the dataframes

ggplot(df_out,aes(x = x1,y = x2,color = y, shape = identifier)) + geom_point() # plot the new df

这会产生：

【讨论】：