Prediction
Finally, we arrive at one of the most important uses of linear regression: prediction. We've trained a model capable of predicting the weight of Olympic swimmers given the data about their height, gender, and year of birth.
Mark Spitz is a nine-time Olympic swimming champion, and he won seven gold medals at the 1972 Olympics. He was born in 1950 and, according to his Wikipedia page, is 183cm tall and weighs 73kg. Let's see what our model predicts as his weight.
Our multiple regression model requires these values to be presented as a matrix form. Each of the parameters needs to be provided in the order in which the model learned the features so that the correct coefficient is applied. After the bias term, our feature vector needs to contain height, gender, and year of birth in the same units as our model was trained:
Our β matrix contains the coefficients for each of these features:
The prediction of our model will be the sum of the products of the β coefficients and features x for...
