Created
December 13, 2017 18:26
-
-
Save fohria/e8d2f23db316c7ca7c55fe05d9617e71 to your computer and use it in GitHub Desktop.
example of multiprocessing usage
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
import pandas as pd | |
from scipy.optimize import minimize | |
import sys | |
from multiprocessing import Pool | |
# local modules and functions imported here | |
sys.path.append('../experiment/simulation') | |
from utils import softmax, autocorrelation | |
class MaxLike(object): | |
def __init__(self, df): | |
self.number_of_actions = 2 | |
self.dataframe = df | |
def neg_log_likelihood(self, alphabeta): | |
df = self.dataframe | |
alpha = alphabeta[0] | |
beta = alphabeta[1] | |
actions, rewards = df.action.values, df['reward'].values | |
prob_log = 0 | |
Q = np.zeros(self.number_of_actions) | |
for action, reward in zip(actions, rewards): | |
Q[action] += alpha * (reward - Q[action]) | |
prob_log += np.log(softmax(Q, beta)[action]) | |
return -2 * prob_log # -2 because reasons | |
def ml_estimation(self, start_parameters): | |
# print("running estimation with parameters: {}".format(start_parameters)) | |
rawr = minimize(self.neg_log_likelihood, start_parameters, | |
method='L-BFGS-B', bounds=((0.01,2), (0.01,5)), | |
options={'disp': False}) | |
return rawr | |
def ml_estimation_multicore(self, start_parameters): | |
""" start_parameters is a list of the form [[0.1, 0.2]..[0.9, 0.9]]""" | |
print("booting up extra core engines to run estimation...") | |
with Pool() as pooler: | |
results = pooler.map(self.ml_estimation, start_parameters) | |
best_bic = 10000 | |
best_index = None | |
for index, result in enumerate(results): | |
lnlike = -result['fun'] | |
bic = 2 * np.log(80) - lnlike | |
if bic < best_bic: | |
best_bic = bic | |
best_index = index | |
return results[best_index] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment