tf = np.arange(1200 * (1 + n_frames * n0),
(nmax + 1) * 1200 * n_frames + 1,
step=1200)
print(f"Loading data took f{time.time() - loadingtimestart:.2f}s. ")
# ----------
# NEWMA RP OPU
startingtime = time.time()
# Pararameters for the NEWMA algorithm
B = 250 # window size
d = n_feat
# Forget factors chosen with heuristic in the paper
big_Lambda, small_lambda = algos.select_optimal_parameters(B)
thres_ff = small_lambda
# Number of random features is set automatically with this criterion
m = int((1 / 4) / (small_lambda + big_Lambda)**2)
m_OPU = 10 * m
print(f'{m_OPU} random features.')
if pyramid:
# Pyramid sketch
nc = 60000
# Opening the OPU
opu_mapping = OPUMap(n_components=nc)
# First encoding
encoder = BinaryThresholdEncoder(threshold_enc=10)
minX, maxX = np.min(traj), np.max(traj)
# parameters of gmm
k = 10
std_mean = 0.11 # the bigger, the more change in means
wishart = 5 # the bigger, the less change in diagonal variances
X, ground_truth = gd.stream_GMM(d=d, n=n, nb_change=nb_change, std_mean=std_mean, concentration_wishart=wishart,
k=k)
# common config
choice_sigma = 'median'
numel = 100
data_sigma_estimate = X[:numel] # data for median trick to estimate sigma
B = args.B # window size
# Newma config
big_Lambda, small_lambda = algos.select_optimal_parameters(B) # forget factors chosen with heuristic in the paper
print('Chose Lambda = {:.4f} and lambda = {:.4f}'.format(big_Lambda, small_lambda))
thres_ff = small_lambda
# number of random features is set automatically with this criterion
m = int((1 / 4) / (small_lambda + big_Lambda) ** 2)
print('Number of RFs: {}'.format(m))
W, sigmasq = feat.generate_frequencies(m, d, data=data_sigma_estimate, choice_sigma=choice_sigma)
print('Start algo ', algo, ' with fixed threshold')
def feat_func(x):
return feat.fourier_feat(x, W)
detector95 = algos.NEWMA(X[0], forget_factor=big_Lambda, forget_factor2=small_lambda, feat_func=feat_func,
adapt_forget_factor=thres_ff)
detector95.apply_to_data(X)
k=k)
# parameters for algorithms
m = 2000
save_res = True
window_size_list = np.linspace(100, 500, num_tests)
N = 3
running_times = []
for i, window_size in enumerate(window_size_list):
window_size = int(window_size)
print('Performing test ', i + 1, ' over ', num_tests, ' window size = ',
window_size)
ff, ff2 = algos.select_optimal_parameters(window_size)
thresh_ff = ff2
if algo == 'ScanB':
print('Scan-B')
ocpobj = algos.ScanB(
X[0],
kernel_func=lambda x, y: feat.gauss_kernel(x, y, np.sqrt(sigmasq)),
window_size=window_size,
nbr_windows=N,
adapt_forget_factor=thresh_ff,
store_values=False)
t = time.time()
ocpobj.apply_to_data(X) # actual computations
time_method = time.time() - t
print('time:', time_method)