Let's say, I have a matrix of probabilities. Each row represents an event and columns represents probabilities of possible outcomes for that event. For example:
4% 30% 35% 31%
34% 18% 21% 27%
25% 19% 21% 35%
29% 14% 25% 32%
I must pick 1, 2, 3 or even 4 outcomes from each row. So, for example if I choose ALL, then the chance to predict every event outcome is 100%.
In general the task is to make as less as possible picks to have the best chance to predict all of the events.
I guess there are some ready solutions, but I'm not sure how to formulate the problem to find them.