def run(self,iterations):
""" Run the Gibbs sampler for the specified number of iterations. """
self.all_U = numpy.zeros((iterations,self.I,self.K))
self.all_V = numpy.zeros((iterations,self.J,self.K))
self.all_tau = numpy.zeros(iterations)
self.all_times = []
self.all_performances = { metric: [] for metric in METRICS }
time_start = time.time()
for it in range(iterations):
# Update the random variables
self.U = update_U_gaussian_volumeprior_nonnegative(
gamma=self.gamma, R=self.R, M=self.M, U=self.U, V=self.V, tau=self.tau)
self.V = update_V_gaussian_gaussian_multivariate(
lamb=self.lamb, R=self.R, M=self.M, U=self.U, tau=self.tau)
self.tau = update_tau_gaussian(
alpha=self.alpha, beta=self.beta, R=self.R, M=self.M, U=self.U, V=self.V)
# Store the draws
self.all_U[it], self.all_V[it] = numpy.copy(self.U), numpy.copy(self.V)
self.all_tau[it] = self.tau
# Print the performance, store performance and time
perf = self.predict_while_running()
for metric in METRICS:
self.all_performances[metric].append(perf[metric])
time_iteration = time.time()
self.all_times.append(time_iteration-time_start)
print "Iteration %s. MSE: %s. R^2: %s. Rp: %s." % (it+1,perf['MSE'],perf['R^2'],perf['Rp'])
def run(self,iterations):
""" Run the Gibbs sampler for the specified number of iterations. """
self.all_U = numpy.zeros((iterations,self.I,self.K))
self.all_V = numpy.zeros((iterations,self.J,self.K))
self.all_tau = numpy.zeros(iterations)
self.all_times = []
self.all_performances = { metric: [] for metric in METRICS }
time_start = time.time()
for it in range(iterations):
# Update the random variables
self.U = update_U_gaussian_gaussian_multivariate(
lamb=self.lamb, R=self.R, M=self.M, V=self.V, tau=self.tau)
self.V = update_V_gaussian_gaussian_multivariate(
lamb=self.lamb, R=self.R, M=self.M, U=self.U, tau=self.tau)
self.tau = update_tau_gaussian(
alpha=self.alpha, beta=self.beta, R=self.R, M=self.M, U=self.U, V=self.V)
# Store the draws
self.all_U[it], self.all_V[it] = numpy.copy(self.U), numpy.copy(self.V)
self.all_tau[it] = self.tau
# Print the performance, store performance and time
perf = self.predict_while_running()
for metric in METRICS:
self.all_performances[metric].append(perf[metric])
time_iteration = time.time()
self.all_times.append(time_iteration-time_start)
print "Iteration %s. MSE: %s. R^2: %s. Rp: %s." % (it+1,perf['MSE'],perf['R^2'],perf['Rp'])
def run(self,iterations):
""" Run the Gibbs sampler for the specified number of iterations. """
self.all_U = numpy.zeros((iterations,self.I,self.K))
self.all_V = numpy.zeros((iterations,self.J,self.K))
self.all_muU = numpy.zeros((iterations,self.K))
self.all_muV = numpy.zeros((iterations,self.K))
self.all_sigmaU = numpy.zeros((iterations,self.K,self.K))
self.all_sigmaV = numpy.zeros((iterations,self.K,self.K))
self.all_tau = numpy.zeros(iterations)
self.all_times = []
self.all_performances = { metric: [] for metric in METRICS }
time_start = time.time()
for it in range(iterations):
# Update the random variables
self.muU, self.sigmaU = update_muU_sigmaU_gaussian_gaussian_wishart(
mu0=self.mu0, beta0=self.beta0, v0=self.v0, W0=self.W0, U=self.U)
self.U = update_U_gaussian_gaussian_wishart(
muU=self.muU, sigmaU=self.sigmaU, R=self.R, M=self.M, V=self.V, tau=self.tau)
self.muV, self.sigmaV = update_muV_sigmaV_gaussian_gaussian_wishart(
mu0=self.mu0, beta0=self.beta0, v0=self.v0, W0=self.W0, V=self.V)
self.V = update_V_gaussian_gaussian_wishart(
muV=self.muV, sigmaV=self.sigmaV, R=self.R, M=self.M, U=self.U, tau=self.tau)
self.tau = update_tau_gaussian(
alpha=self.alpha, beta=self.beta, R=self.R, M=self.M, U=self.U, V=self.V)
# Store the draws
self.all_U[it], self.all_V[it] = numpy.copy(self.U), numpy.copy(self.V)
self.all_muU[it], self.all_sigmaU[it] = numpy.copy(self.muU), numpy.copy(self.sigmaU)
self.all_muV[it], self.all_sigmaV[it] = numpy.copy(self.muV), numpy.copy(self.sigmaV)
self.all_tau[it] = self.tau
# Print the performance, store performance and time
perf = self.predict_while_running()
for metric in METRICS:
self.all_performances[metric].append(perf[metric])
time_iteration = time.time()
self.all_times.append(time_iteration-time_start)
print "Iteration %s. MSE: %s. R^2: %s. Rp: %s." % (it+1,perf['MSE'],perf['R^2'],perf['Rp'])