rewardplot.py 文件源码

def plot_avg_regret(policy):

    """Plot average regret with respect to time.

        Parameters
        ----------
        policy: bandit object
            The bandit algorithm you want to evaluate.
    """

    avg_reward = calculate_avg_reward(policy)
    points = sorted(six.viewitems(avg_reward), key=lambda x: x[0])
    x, y = zip(*points)
    plt.plot(x, [1 - reward for reward in y], 'r-', label="average regret")
    plt.xlabel('time')
    plt.ylabel('avg regret')
    plt.legend()
    plt.title("Average Regret with respect to Time")