Closed raul-singh closed 2 years ago
@RaulSingh-7 Based on the changes we can see that now the tests for the clairvoyant learner is failing because it's predict function assumes that the class parameters is equal for each product. Hence we need to modify that aswell before merging this.
I see. I'll look into it and try to solve this as soon as possible.
Ok, so @barskern the main problem I found is that the clairvoyant learner and the function calculate_aggregated_budget_value
do their computation on the assumption that every class has a single alpha function. To solve this I'll have to modify the ClairvoyantLearner
and the calculate_aggregated_budget_value
function.
@baskern I updated the ClairvoyantLearner
and calculate_aggregated_budget_value
. It was harder than expected but now the tests go flawlessly.
Oups, seems like I didnt format it correctly, would you mind pulling the changes, formatting locally and pushing again? 😇 @RaulSingh-7
I made some small modifications to the alpha functions as we discussed yesterday, so now each individual class has an alpha function for each product. Only thing probably the
ClairvoyantLearner
is gonna break because of these changes.