I have a probability problem. Let's say I have an event sale and I have a 1 million population for whom I was to predicted their probability to sale. I use a machine learning model to learn from a sample population and then predict for each person in 1 M population their probability to sale.
Is it mathematically ok to say that if I average the predicted probabilities overall the whole population of 1 million,and let's say it comes out to be 0.3,
1) on average 30% of the population is going to sale
2) each random person has a 30% probability to sale
if it is not right to interpret it like this. Then what should be the interpretation?
Those are both reasonable interpretations of your average.