Unlocking the Power of Prediction Intervals in R: A Practical Guide

[This article was first published on Steve's Data Tips and Tricks, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Introduction

Prediction intervals are a powerful tool for understanding the uncertainty of your predictions. They allow you to specify a range of values within which you are confident that the true value will fall. This can be useful for many tasks, such as setting realistic goals, making informed decisions, and communicating your findings to others.

In this blog post, we will show you how to create a prediction interval in R using the mtcars dataset. The mtcars dataset is a built-in dataset in R that contains information about fuel economy, weight, displacement, and other characteristics of 32 cars.

Creating a Prediction Interval

To create a prediction interval in R, we can use the predict() function. The predict() function takes a fitted model and a new dataset as input and returns the predicted values for the new dataset.

We can also use the predict() function to calculate prediction intervals. To do this, we need to specify the interval argument. The interval argument can take two values: confidence and prediction.

A confidence interval is the range of values within which we are confident that the true mean of the population will fall. A prediction interval is the range of values within which we are confident that the true value of a new observation will fall.

To create a prediction interval for the mpg variable in the mtcars dataset, we can use the following code:

# Fit a linear model
model <- lm(mpg ~ disp, data = mtcars)

# Create a prediction interval
prediction_intervals <- predict(
  model, 
  newdata = mtcars, 
  interval = "prediction", 
  level = 0.95
  )

# Print the prediction interval
head(prediction_intervals)
                       fit       lwr      upr
Mazda RX4         23.00544 16.227868 29.78300
Mazda RX4 Wag     23.00544 16.227868 29.78300
Datsun 710        25.14862 18.302683 31.99456
Hornet 4 Drive    18.96635 12.217933 25.71477
Hornet Sportabout 14.76241  7.905308 21.61952
Valiant           20.32645 13.582915 27.06999

The prediction interval shows that we are 95% confident that the true mpg value for a new car with a given displacement will fall within the range specified by the lwr and upr columns.

Visualize

First lets bind the data together with cbind()

full_res <- cbind(mtcars, prediction_intervals)

head(full_res)
                   mpg cyl disp  hp drat    wt  qsec vs am gear carb      fit
Mazda RX4         21.0   6  160 110 3.90 2.620 16.46  0  1    4    4 23.00544
Mazda RX4 Wag     21.0   6  160 110 3.90 2.875 17.02  0  1    4    4 23.00544
Datsun 710        22.8   4  108  93 3.85 2.320 18.61  1  1    4    1 25.14862
Hornet 4 Drive    21.4   6  258 110 3.08 3.215 19.44  1  0    3    1 18.96635
Hornet Sportabout 18.7   8  360 175 3.15 3.440 17.02  0  0    3    2 14.76241
Valiant           18.1   6  225 105 2.76 3.460 20.22  1  0    3    1 20.32645
                        lwr      upr
Mazda RX4         16.227868 29.78300
Mazda RX4 Wag     16.227868 29.78300
Datsun 710        18.302683 31.99456
Hornet 4 Drive    12.217933 25.71477
Hornet Sportabout  7.905308 21.61952
Valiant           13.582915 27.06999

Now let’s plot the actual, the fitted and the prediction confidence bands.

library(ggplot2)

full_res |>
  ggplot(aes(x = disp, y = mpg)) +
  geom_point() +
  geom_point(aes(y = fit), col = "steelblue", size = 2.5) +
  geom_line(aes(y = fit)) +
  geom_line(aes(y = lwr), linetype = "dashed", col = "red") +
  geom_line(aes(y = upr), linetype = "dashed", col = "red") +
  theme_minimal() +
  labs(
    title = "mpg ~ disp, data = mtcars",
    subtitle = "With Prediction Intervals"
  )

Above we are capturing the prediction interval which gives us the uncertainty around a single point, whereas the confidence interval gives us the uncertainty around the mean predicted values. This means that the prediction interval will always be wider than the confidence interval for the same value.

Trying It Out Yourself

Now it’s your turn to try out creating a prediction interval in R. Here are some ideas:

  • Try creating a prediction interval for a different variable in the mtcars dataset, such as wt or hp.
  • Try creating a prediction interval for a variable in a different dataset.
  • Try creating a prediction interval for a more complex model, such as a multiple linear regression model or a logistic regression model.

Conclusion

Creating prediction intervals in R is a straightforward process. By using the predict() function, you can easily calculate prediction intervals for any fitted model and any new dataset. This can be a valuable tool for understanding the uncertainty of your predictions and making more informed decisions.

To leave a comment for the author, please follow the link and comment on their blog: Steve's Data Tips and Tricks.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)