return pred_mean, pred_var
if __name__ == '__main__':
'''
If your implementations are correct, running
python problem.py
inside the Bayesian Regression directory will, for each sigma in sigmas_to-test generates plots
'''
np.random.seed(46134)
actual_weights = np.matrix([[0.3], [0.5]])
data_size = 40
noise = {"mean": 0, "var": 0.2**2}
likelihood_var = noise["var"]
xtrain, ytrain = support_code.generate_data(data_size, noise,
actual_weights)
#Question (b)
sigmas_to_test = [1 / 2, 1 / (2**5), 1 / (2**10)]
for sigma_squared in sigmas_to_test:
prior = {
"mean": np.matrix([[0], [0]]),
"var": matlib.eye(2) * sigma_squared
}
support_code.make_plots(actual_weights, xtrain, ytrain, likelihood_var,
prior, likelihood_func, get_posterior_params,
get_predictive_params)
# TO DO
return predMean, predVar
if __name__ == '__main__':
'''
If your implementations are correct, running
python problem.py
inside the Bayesian Regression directory will, for each sigma in sigmas_to-test generates plots
'''
np.random.seed(46134)
actual_weights = np.matrix([[0.3], [0.5]])
dataSize = 40
noise = {"mean": 0, "var": 0.2**2}
likelihood_var = noise["var"]
xtrain, ytrain = support_code.generateData(dataSize, noise, actual_weights)
#Question (b)
sigmas_to_test = [1 / 2, 1 / (2**5), 1 / (2**10)]
for sigma_squared in sigmas_to_test:
prior = {
"mean": np.matrix([[0], [0]]),
"var": matlib.eye(2) * sigma_squared
}
support_code.make_plots(actual_weights, xtrain, ytrain, likelihood_var,
prior, likelihoodFunc, getPosteriorParams,
getPredictiveParams)