Question

我有以下情节：

在那里，我用Serial_number因子的每个级别绘制数据框的拟合值。我想摆脱所有错误的级别，意思是，只有带有608004650的Serial_number出现。

代码：

> head(APD)
# A tibble: 6 x 7
  Serial_number    Lot  Wafer Amplification Voltage Dark_current     pred
         <fctr> <fctr> <fctr>         <dbl>   <dbl>        <dbl>    <dbl>
1     608004648      6    608       111.997 379.980     0.386364 1.550965
2     608004648      6    608       123.673 381.968     0.381323 1.572166
3     608004648      6    608       137.701 383.979     0.411581 1.594672
4     608004648      6    608       154.514 385.973     0.460648 1.618040
5     608004648      6    608       175.331 387.980     0.506632 1.642617
6     608004648      6    608       201.379 389.968     0.554607 1.668009

fit<- lmer(log(log(Amplification)) ~ poly(Voltage, 2) + (poly(Voltage, 2) | Serial_number), data = APD)
APD$pred<- fitted(fit)

ggplot(APD, aes(x = Voltage, y = log(log(Amplification)), colour = Serial_number==608004650)) +
  geom_point(size=3) +
  geom_line(aes(y = fitted(fit)),size=1)

编辑：缺少代码行来理解数据构造：

Limit data ranges:

APD.frame$Threshold<- "No"
APD.frame$Threshold[APD.frame$Amplification <= 150]<- "Yes"
APD.frame <- within(APD.frame,{
  Threshold <- factor(Threshold)
})


  APDmin<- APD.frame %>% group_by(Serial_number, Threshold)  %>% top_n(n, Amplification)
  APDmin<- APDmin[!APDmin$Threshold == "No", ]
  APDmax<- APD.frame %>% group_by(Serial_number, Threshold)  %>% top_n(-n, Amplification)
  APDmax<- APDmax[!APDmax$Threshold == "Yes", ]
  APD<- rbind(APDmin, APDmax)
  APD<- with(APD, APD[order(Serial_number, Amplification),])
  APD$Threshold<- NULL

最小可重复的示例（34个代码行+部分数据（~270行））：https://ufile.io/345np

编辑：当我不删除Threshold列并应用相同的绘图命令时，我现在得到：

str(APD.frame <- read.table("4_APDs_irradiated.txt", header = TRUE, sep = "", dec = "."))

APD.frame <- within(APD.frame,{
  Serial_number <- factor(Serial_number)
  Irradiated <- factor(Irradiated, levels = c(0, 1), labels = c("no", "yes"))
})
APD.frame$Irradiated<- NULL

library(dplyr)
library(lme4)

APD.frame$Threshold<- "No"
APD.frame$Threshold[APD.frame$Amplification <= 150]<- "Yes"

n = 3
APDmin<- APD.frame %>% group_by(Serial_number, Threshold)  %>% top_n(n, Amplification)
APDmin<- APDmin[!APDmin$Threshold == "No", ]
APDmax<- APD.frame %>% group_by(Serial_number, Threshold)  %>% top_n(-n, Amplification)
APDmax<- APDmax[!APDmax$Threshold == "Yes", ]
APD<- rbind(APDmin, APDmax)
APD<- with(APD, APD[order(Serial_number, Amplification),])

summary(fit<- lmer(log(log(Amplification)) ~ poly(Voltage, 2) + (poly(Voltage, 2) | Serial_number), data = APD))

APD$pred<- fitted(fit)

APD %>% filter(Serial_number == "912009895") %>% 
  ggplot(aes(x = Voltage, y = log(log(Amplification)))) +
  geom_line(aes(y = fitted(fit), size=2)) 
Error: Aesthetics must be either length 1 or the same as the data (6): y, size, x

Answer 1

我有一个解决方法：

APD$SN<- "No"
APD$SN[APD$Serial_number == 912009895]<- "Yes"
One_APD<- subset(APD, SN=="Yes")
ggplot(One_APD, aes(x = Voltage, y = log(log(Amplification)))) +
  geom_point(size=3, colour="blue") +
  geom_line(aes(y = pred, size=0.5, alpha=0.2, colour="red")) + 
  theme(legend.position="none")

虽然它有效，但我感到惊讶的是，并没有像＃34;掉落等级错误＆＃34; ..

dgpe ggplot中的错误级别

1 个答案: