CODEWITHSUNDEEP
rplotregression
Yoav
Yoav

Reputation: 1

How to fit regression lines per group ( each group contains two different factors/columns)?

I have managed to plot a generalized line, but it is not relevant to my needs. I want to plot line per each treatment and ID and to get the regression values. pic attached here: enter image description here

I am relatively new with R and couldn't find a proper answer online. Thanks

This is my data:

enter image description here

I want to plot the regression lines for each combination of treatment (0/4/8)* and patient type.

the code that I have written is:

plot(ion_chlorophyll$NA,ion_chlorophyll$CHL)
plot(ion_chlorophyll$NA,ion_chlorophyll$CHL, pch = 1,cex = 1, col = "blue", main = "NA relationship", xlab ="Na", ylab ="CHL")

Upvotes: 0

Views: 1748

Answers (2)

xyz123
xyz123

Reputation: 651

if you want to get multiple regression lines try using xyplot. Experiment with the iris dataset. 

library(lattice)

xyplot(Sepal.Length ~ Sepal.Width, data = iris, pch = 16, type = c("p", "g", "r"), groups = Species, auto.key = TRUE)

If you just have three factors contributing to your 2D plot, then it shouldn't be too much trouble to get the regression lines this way: 

model1 <- lm(iris$Sepal.Length[iris$Species == "setosa"] ~ iris$Sepal.Width[iris$Species == "setosa"])

model2 <- lm(iris$Sepal.Length[iris$Species == "versicolor"] ~ iris$Sepal.Width[iris$Species == "versicolor"])

model3 <- lm(iris$Sepal.Length[iris$Species == "virginica"] ~ iris$Sepal.Width[iris$Species == "virginica"])

And do a summary() of each of the models to get the equations of the regression lines.

I think you should do something like this.

enter image description here

Upvotes: 1

Karthik Arumugham
Karthik Arumugham

Reputation: 1350

You can do it with ggplot2 if you just want to plot. You can use dplyr, tidy and broom to get the predictions from the model. If df is your dataframe with columns - treatment, patient_type, Leaf_NA, Total_Chl,

Just plot

library(ggplot2)
p <- df %>% ggplot(aes(Leaf_NA, Total_Chl)) +
    facet_grid(treatment ~ patient_type) +
    stat_smooth(method='lm')
p

Get model predictions

library(dplyr) #for group_by(), inner_join(), mutate()
library(tidyr) #for nest(), unnest()
library(broom) #for augment()

model <- df %>%
    group_by(treatment, patient_type) %>%
    do(fit = lm(Total_Chl ~ Leaf_NA, data = .))

df %>%
    group_by(treatment, patient_type) %>%
    nest() %>%
    inner_join(model, .) %>%
    mutate(pred = list(augment(fit))) %>%
    unnest(pred)

Upvotes: 0

Related Questions