In the Introduction to Linear Regression I have shown how is possible to use one single feature input – like the average season temperature – to predict another variable – like the wine price. Using an approach called Linear Regression.
But often we have many different features which we could use as inputs. As we have seen in the introduction example, other features – such as the amount of rain in the season – were available.
You might go in to the historical data set and see that two years had very similar average temperature but one was much more rainy than the other. Has this an effect on the wine quality and therefore on its price? And if yes, which kind of effect, positive or negative? Small or big?
In this section we want to use multiple features (“the independent variables”) to predict the wine price (“the dependent variable”). In particular, in this higher dimensional space, we want to fit some kind of function that models the relationship between these inputs and the output. Continue reading “Multiple Linear Regression”