extract_interval_mean_at_a_time_step()

home > kero > Documentation

Extract the mean number of actions over an interval, which is given by the ration of the number of that action to the number of actions in the interval. The function extracts the value at a single time step, and reorganize the data to be based on the action on the highest level rather than the  member.

kero.multib.nDnet.py
class Network_processor:
  extract_interval_mean_at_a_time_step(self,interval_mean_history,time_step=None):
    return interval_mean_at_t

Arguments/Return

interval_mean_history Data of the format specified by the return value of  compute_interval_mean().
time_step Integer. The time step at which the mean values are extracted from interval_mean_history. If set to None, the last time step is chosen.

Default=None

return interval_mean_at_t Dictionary:
– key: (Integer) action index
– values: Dictionary
+ key: (String) network member
+ values: list, [x] where x is the mean value of the number of an action specified by the action index over some time steps.Example output: {0: {‘1’: [0.25], ‘2’: [0.5], ‘3’: [0.5], ‘4’: [0.25]}, 1: {‘1’: [0.75], ‘2’: [0.5], ‘3’: [0.5], ‘4’: [0.75]}}

Example Usage

See here.

kero version: 0.5.1 and above