When we apply a model to data, we have numerous explanatory variables. With these variables, we would like to explain the dependent/response variable in order to model it. It is possible that these variables represent the same quantity, as they share the same part of the variance. So, to choose which variables to use, we need to apply model selection methods, the aim being to find the simplest model, with the smallest error and the fewest explanatory variables.
To the video