Random Test Dataset With More Values Of One Type

I have a set with y values of 0 or 1.

most of my data has 0 value, so when splitting the data to train and test 80 -20 respectively, most of the test data has 0 y value.

I want that the randomly chosen test data will have more data with 1 has y vlaues, how can it be done?

Answer 1

[the cyclist]

One simple approach would be:

1. split your dataset into the y=0 set and the y = 1 set
2. do the 80/20 training/test split on the y=0 and y=1 separately
3. combine the two training sets and the two test sets

Bear in mind that highly imbalanced datasets like this have pitfalls for analysis. It's too much to describe here, but if you search keywords like imbalanced dataset machine learning, you'll be able to read about the problems, and some potential solutions.