''' Run the experiment. '''
R, M = load_gdsc_ic50_integer()
model_class = BMF_Gaussian_Exponential
settings = {
'R': R,
'M': M,
'K': 20,
'hyperparameters': {
'alpha': 1.,
'beta': 1.,
'lamb': 0.1
},
'init': 'random',
'iterations': 200,
}
fout_performances = './results/performances_gaussian_exponential.txt'
fout_times = './results/times_gaussian_exponential.txt'
repeats = 10
performances, times = measure_convergence_time(repeats, model_class, settings,
fout_performances, fout_times)
''' Plot the times, and performance vs iterations. '''
plt.figure()
plt.title("Performance against average time")
plt.plot(times, performances)
plt.ylim(0, 2000)
plt.figure()
plt.title("Performance against iteration")
plt.plot(performances)
plt.ylim(0, 2000)
''' Run the experiment. '''
R, M = load_movielens_100K()
model_class = BMF_Gaussian_Exponential_ARD
settings = {
'R': R,
'M': M,
'K': 20,
'hyperparameters': { 'alpha':1., 'beta':1., 'alpha0':1., 'beta0':1. },
'init': 'random',
'iterations': 200,
}
fout_performances = './results/performances_gaussian_exponential_ard.txt'
fout_times = './results/times_gaussian_exponential_ard.txt'
repeats = 10
performances, times = measure_convergence_time(
repeats, model_class, settings, fout_performances, fout_times)
''' Plot the times, and performance vs iterations. '''
plt.figure()
plt.title("Performance against average time")
plt.plot(times, performances)
plt.ylim(0,2000)
plt.figure()
plt.title("Performance against iteration")
plt.plot(performances)
plt.ylim(0,2000)
