Chronological Order in Boosting
Let’s say you’re analyzing a Netflix users dataset that was collected over a period of 10 years.
You are training a boosting algorithm on the data to predict whether a user will trust the website, indicated by them entering their credit card info for a trial period.
When training your model, would it be a good idea to separate the user groups based on the year they became a Netflix member? Why or why not?
Next question: Check Normality.....
Loading comments