If we have e.g 10 features it is difficult to understand the data shape and its linearity until we run the linear and non-linear model and compear the results.Do you have any suggestions ?
Since a linear model will be really cheap to fit, it is a good baseline to start with before running a more complex model. Making a statistical descriptive analysis as done in the first module can help as well: checking the pair plot can give a little insight.
1 Like