The “predict” Function in R

• Package: base

• Purpose: Generate predictions from a fitted model.

• General class: Method

• Required argument(s):

• object: A fitted model object (e.g., from lm, glm, randomForest).

• newdata: Data frame or matrix containing the new data for which predictions are to be made.

• Notable optional arguments:

• type: Type of prediction (e.g., “response” for the fitted values, “prob” for probabilities, depending on the model).

• se.fit: Boolean indicating if standard errors for predictions should be returned (specific to some models).

• Example:

• # Load the required library
  library(MASS)
  
  # Example dataset
  data(Boston)
  
  # Fit a linear model
  model <- lm(medv ~ ., data = Boston)
  
  # New data for prediction
  new_data <- Boston[1:5, ]
  
  # Generate predictions
  predictions <- predict(model, newdata = new_data, type = "response")
  
  # Print predictions
  print(predictions)

• In this example, the predict function is used to generate predictions from a linear model fitted to the Boston dataset. The newdata argument specifies the data for which predictions are to be made, and type = "response" indicates that we want the predicted values of the response variable. The function outputs the predictions for the specified new data.