python

如何快速建立数字一样的array

>>> R = 4 * [400.]
>>> R

[400.0, 400.0, 400.0, 400.0]

依次改动np.array中的值

import numpy as np

def compute_EA(RA, RB):
    '''
        compute the expected probability of player A to win in a game with player B.
        Input:
            RA: the rating of player A, a float scalar value
            RB: the rating of player B, a float scalar value
        Output:
            EA: the expected probability of A wins, a float scalar value between 0 and 1.
    '''
    #########################################
    ## INSERT YOUR CODE HERE
    EA = 1 / (1 + pow(10, ((RB - RA) / 400)))

    #########################################
    return EA

#--------------------------
def update_RA(RA, SA, EA, K = 16.):
    '''
        compute the new rating of player A after playing a game.
        Input:
            RA: the current rating of player A, a float scalar value
            SA: the game result of player A, a float scalar value.
                if A wins in a game, SA = 1;if A loses, SA =0.
            EA: the expected probability of player A to win in the game, a float scalar between 0 and 1.
             K: k-factor, a contant number which controls how fast to correct the ratings
        Output:
            RA_new: the new rating of player A, a float scalar value
    '''
    #########################################
    ## INSERT YOUR CODE HERE
    RA_new = RA + K * (SA - EA)
    #########################################
    return RA_new


#--------------------------
def elo_rating(W, n_player, K= 16.):
    '''
        An implementation of Elo rating algorithm, which was used in facemash.
        Given a collection of game results W, compute the Elo rating scores of all the players.
        Input:
                W: (wins) game results, a numpy matrix of shape (n_game,2), dtype as integers. If player i wins player j in the k-th game, W[k][0] = i, W[k][1] = j.
                n_player: the number of players to rate, an integer scalar.
                K: k-factor, a contant number which controls how fast to correct the ratings
        Output:
                R: the Elo rating scores,  a python array of float values, such as [1000., 200., 500.], of length num_players
    '''

    # initialize the ratings of all players with 400
    R = n_player * [400.]

    # for each game, update the ratings
    for (A, B) in W:
        # the game result: player A (win), player B (loss)
        # A is the index of player A, B is the index of player B

        # update player A's rating
        ##############################
        # INSERT YOUR CODE HERE
        RA = R[A] 
        # player A's rating
        RB = R[B] 
        # player B's rating
        EA = compute_EA(RA, RB) 
        # the expected probability of (player A wins the game)
        R[A] = update_RA(RA, 1., EA, K) 
        # update A's rating
        ##############################

        # update player B's rating
        ##############################
        # INSERT YOUR CODE HERE
        #EB = compute_EA(RB, RA) # the expected probability of (player B wins the game)
        EB = 1. - EA # the expected probability of (player B wins the game)
        R[B] = update_RA(RB, 0., EB, K) # update B's rating
        ##############################
    return R

W = np.array([[1,0],[2,4],[5,3],[1,2],[3,4],[5,3]])
n_player = 6

R = elo_rating(W, n_player, K= 16.)

R