Yoonki Jeong

Seems 'prob_leave_bandit' in 'state_transition' not used.

1

For both RecoGym environment (RecoEnv0, RecoEnv1), they initialize 'state_transition' using default arguments passed through config. It seems that the second row, which correspond to bandit, use 'self.config.prob_leave_organic' instead of 'self.config.prob_leave_bandit'....

"relevent_num" should be minimum value between (k, # ground truth). Isn't "np.cumsum" supposed to be "np.array"? https://github.com/wubinzzu/NeuRec/blob/91ef5e79b8b8e420da65a52d774406114db23e0d/evaluator/backend/python/metric.py#L39

# Possible Issue In bandit feedback, n_actions are set as ```int(self.action.max() + 1)```, which doesn't raise any error in above code, assuming that logs generated by ```policy``` covered all possible...