Making predictions

Okey, so let's assume that we have a linear model which seems to fit our standards. What can we do with it?

The model quantifies the relationship between the explanatory variable(s) and the dependent variable. The model can also be used for predicting the dependent variable based on new observations of the explanatory variable(s).

In R, predicting can be done using the predict() function. (see ?predict). The first argument of predict is a model object and the argument newdata (a data.frame) can be used to make predictions based on new observations. One or more columns of newdata should have the same name as the explanatory variables in the model object.

This exercise is part of the course

Helsinki Open Data Science

View Course

Exercise instructions

Create object m and print out a summary of the model
Create object new_attitudes
Adjust the code: Create a new data frame with a column named 'attitude' holding the new attitudes defined in new_attitudes
Print out the new data frame
predict() the new student's exam points based on their attitudes, using the newdata argument

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# learning2014 is available

# Create model object m
m <- lm(points ~ attitude, data = learning2014)

# print out a summary of the model


# New observations
new_attitudes <- c("Mia" = 3.8, "Mike"= 4.4, "Riikka" = 2.2, "Pekka" = 2.9)
new_data <- data.frame(attitude = "change me!")

# Print out the new data


# Predict the new students exam points based on attitude
predict(m, newdata = "change me!")

Edit and Run Code