So far, we have defined the problem and designed the approach. We explored the data and studied the patterns across a variety of parameters captured through the sensors. We then engineered the data and created a couple of features that depict the day-level activities in an enriched dimension. We now have the data with multiple predictors and the dependent variable outcome (created by taking a lead operation on the flag, that is, indicator whether there was a power outage the next day).
We are challenged with the vanilla classification problem with a binary outcome, that is, 1 and O.
Note
As a part of the modeling exercise, we need to explore in depth the variables for the classification model, study correlation, multicollinearity, and other tests, and so on Covering the entire journey of getting data aware for the predictive model building exercise would be out of scope for the chapter. It is highly recommended to execute all the required checks before...