btb.selection.recent module

class btb.selection.recent.RecentKReward(choices, k=2)

Bases: btb.selection.ucb1.UCB1

Recent K reward selector

Parameters

k (int) – number of best scores to consider

compute_rewards(scores)

Retain the K most recent scores, and replace the rest with zeros

select(choice_scores)

Use the top k learner’s scores for usage in rewards for the bandit calculation

class btb.selection.recent.RecentKVelocity(choices, k=2)

Bases: btb.selection.recent.RecentKReward

Recent K velocity selector

compute_rewards(scores)

Compute the velocity of thte k+1 most recent scores.

The velocity is the average distance between scores. Return a list with those k velocities padded out with zeros so that the count remains the same.