def main():
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1, projection='3d')
param_values1 = list(range(1, 401))
param_values2 = list(range(1, 31))
file_name = '/media/nickyz/Data/scriptie_data/results/kalman_filter/trilateration/results_trilateration_{}_{}_{}.csv'
results = []
for i, param_value1 in enumerate(param_values1):
for param_value2 in param_values2:
try:
results[i].append(
get_performance.main(
file_name.format(TIME_WINDOW, param_value1,
param_value2), 'euc')[0])
except IndexError:
results.append([
get_performance.main(
file_name.format(TIME_WINDOW, param_value1,
param_value2), 'euc')[0]
])
min_tup = np.unravel_index(np.argmin(results), np.array(results).shape)
print(np.min(results), np.max(results), param_values1[min_tup[0]],
param_values2[min_tup[1]])
def main():
param_values1 = [float(i) for i in range(31, 51)]
param_values2 = [float(i) for i in range(40001, 50001)]
file_name = '/media/nickyz/Data/scriptie_data/results/dbl_irr_exp_filter/grid/sum/results_grid_normal_signal_sum_{}_{}_{}.csv'
print('{')
for window in TIME_WINDOWS:
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1, projection='3d')
results = []
min_val = 10000000
min_tup = None
for i, param_value1 in enumerate(param_values1):
for param_value2 in param_values2:
new_val = get_performance.main(
file_name.format(window, param_value1, param_value2),
'euc')[0]
if new_val < min_val:
min_val = new_val
min_tup = (param_value1, param_value2)
# try:
# results[i].append(
# get_performance.main(file_name.format(window, param_value1, param_value2), 'euc')[0]
# )
# except IndexError:
# results.append([get_performance.main(file_name.format(window, param_value1, param_value2), 'euc')[0]])
# min_tup = np.unravel_index(np.argmin(results), np.array(results).shape)
# print(' {}: ({}, {}),'.format(window, param_values1[min_tup[0]], param_values2[min_tup[1]]))
print(' {}: ({}, {}),'.format(window, min_tup[0],
min_tup[1]))
print(min_val)
print(' }')
def main():
param_values = range(1, 36)
file_names = [
('grid/std_estimation/results_grid_normal_signal_{}_{}.csv', 'Signal'),
('grid/std_estimation/results_grid_normal_frequency_{}_{}.csv',
'Frequency'),
('grid/std_estimation/results_grid_normal_signal_sum_{}_{}.csv',
'Signal Sum'),
]
results = [[] for _ in file_names]
for param_value in param_values:
for i, file_name in enumerate(file_names):
results[i].append(
get_performance.main(
file_name[0].format(TIME_WINDOW, param_value), 'euc')[0])
for i, file_name in enumerate(file_names):
plt.plot(param_values, results[i])
plt.xlabel('Variance')
plt.ylabel('Average Euclidean Distance')
plt.title(
'Parameter plot to determine the standard deviation using the {}\n normal grid method and a time window of {}'
.format(file_name[1], TIME_WINDOW))
plt.show()
def get_latex_table():
results = {}
for window in TIME_WINDOWS:
base_files = get_files(window)
for i, base_file in enumerate(base_files):
dist, std = get_performance.main(base_file[0], 'euc')
try:
results[base_file[1]].append(dist)
except KeyError:
results[base_file[1]] = [dist]
print('', end=" ")
for window in TIME_WINDOWS:
print('& ' + str(window), end=" ")
print('\\\\')
print('\\hline')
for key in results:
print(key, end=" ")
for value in results[key]:
print('& ' + str(round(value, 3)), end=" ")
print('\\\\')
print('\\hline')
def main():
param_values = range(0, 101)
param_values = [param_value / 100 for param_value in param_values]
file_names = [
('grid/alpha_estimation_signal/results_grid_normal_signal_{}_{}.csv',
'Signal'),
('grid/alpha_estimation_frequency/results_grid_normal_frequency_{}_{}.csv',
'Frequency'),
('grid/alpha_estimation_signal_sum/results_grid_normal_signal_sum_{}_{}.csv',
'Signal Sum'),
]
results = [[] for _ in file_names]
for param_value in param_values:
for i, file_name in enumerate(file_names):
results[i].append(
get_performance.main(
file_name[0].format(TIME_WINDOW, param_value), 'euc')[0])
print(np.argmin(results[0]), np.argmin(results[1]), np.argmin(results[2]))
for i, file_name in enumerate(file_names):
plt.plot(param_values, results[i])
plt.xlabel('Alpha')
plt.ylabel('Average Euclidean Distance')
plt.title(
'Parameter plot to determine the alpha value using the {}\n normal grid method and a time window of {}'
.format(file_name[1], TIME_WINDOW))
plt.show()
def main():
param_values = [i for i in range(0, 41)]
# results = []
#
# for param_value in param_values:
# results.append(
# get_performance.main(NAME.format(TIME_WINDOW, param_value), 'euc')[0]
# )
#
# # print(param_values[np.argmin(results)])
#
#
# plt.plot(param_values, results)
# plt.xlabel('Gamma')
# plt.ylabel('Average Euclidean Distance')
# plt.title('Parameter plot to determine the gamma for the trilateration method\n and a time window of {}'.format(TIME_WINDOW))
# plt.show()
#
# results = []
#
# for param_value in param_values:
# results.append(
# get_performance.main('shift_' + NAME.format(TIME_WINDOW, param_value), 'euc')[0]
# )
#
# # print(param_values[np.argmin(results)])
#
#
# plt.plot(param_values, results)
# plt.xlabel('Gamma')
# plt.ylabel('Average Euclidean Distance')
# plt.title('Parameter plot to determine the gamma for the trilateration method\n and a time window of {}'.format(TIME_WINDOW))
# plt.show()
print('{')
for window in TIME_WINDOWS:
results = []
for param_value in param_values:
results.append(
get_performance.main(NAME.format(window, param_value), 'euc')[0]
)
print(' {}: {},'.format(window, param_values[np.argmin(results)]))
print(' }')
def main():
param_values = [i / 100 for i in range(0, 301, 5)]
file_name = 'trilateration/results_trilateration_{}_{}.csv'
results = []
for param_value in param_values:
results.append(
get_performance.main(file_name.format(TIME_WINDOW, param_value),
'euc')[0])
print(param_values[np.argmin(results)])
plt.plot(param_values, results)
plt.xlabel('Gamma')
plt.ylabel('Average Euclidean Distance')
plt.title(
'Parameter plot to determine the gamma for the trilateration method\n and a time window of {}'
.format(TIME_WINDOW))
plt.show()
def main():
N = len(TIME_WINDOWS)
results = []
labels = ['Euclidean Distance', 'Method', 'Time Window']
methods = []
ps = []
prev_length = 0
for window in TIME_WINDOWS:
base_files = get_files(window)
methods = []
for i, base_file in enumerate(base_files):
dists = get_performance.main(base_file[0], 'euc', False)
methods.append(base_file[2])
# len_dif =
# prev_length = len()
for val in dists:
results.append([val, base_file[1], window])
# tri_shortness = len(part_results[0]) - len(part_results[-1])
#
# while tri_shortness > 0:
# part_results[-1].append(None)
# tri_shortness -= 1
#
# part_results.append([window for _ in part_results[0]])
#
# results += np.array(part_results).T.tolist()
df = pd.DataFrame.from_records(results, columns=labels)
g = sns.boxplot(x='Method',
y='Euclidean Distance',
hue='Time Window',
data=df)
g.set_xticklabels(labels=methods, rotation=30)
axes = g.axes
axes.set_ylim(0, 40)
plt.show()
def main():
param_values = [i / 10 for i in range(1, 1501)]
file_name = 'irr_exp_filter/trilateration/results_trilateration_{}_{}.csv'
print('{')
for window in TIME_WINDOWS:
results = []
params = param_values
for param_value in param_values:
try:
results.append(
get_performance.main(file_name.format(window, param_value),
'euc')[0])
except:
if window == 60:
params = [i / 10 for i in range(1, 501)]
else:
params = [i / 10 for i in range(1, 301)]
break
print(' {}: {},'.format(window,
param_values[np.argmin(results)]))
print(' }')
def main():
results = {}
for time_window in TIME_WINDOWS:
results[time_window] = {}
filter_files = get_files(time_window)
for i, filter_file in enumerate(filter_files):
dist, std = get_performance.main(filter_file[0], 'euc')
try:
results[time_window][filter_file[1]][filter_file[2]] = dist
except:
results[time_window][filter_file[1]] = {filter_file[2]: dist}
f, ((ax1, ax2, ax3, ax4), (ax5, ax6, ax7,
ax8)) = plt.subplots(2, 4, sharey='row')
f.delaxes(ax8)
filter_plot_map = {
'Exponential': ax1,
'Irregular Exponential': ax2,
'Double Exponential': ax3,
'Median': ax4,
'Gauss Median': ax5,
'Shift Median': ax6,
'Kalman': ax7,
}
x_ticks = [
'proxsig', 'proxfreq', 'proxsum', 'gridsig', 'gridfreq', 'gridsum',
'normsumsig', 'normsumfreq', 'normsumsum', 'normhadsig', 'normhadfreq',
'normhadsum', 'trilat'
]
for time_window in results:
for filter_method in filter_methods:
subplot = filter_plot_map[filter_method]
xs = []
ys = []
filter_results = results[time_window][filter_method]
for base_method in x_ticks:
xs.append(X_VALS[base_method])
ys.append(filter_results[base_method])
subplot.scatter(xs, ys, c=COLORS[time_window])
subplot.set_xticks(xs)
subplot.set_xticklabels(x_ticks)
subplot.set_title(filter_method)
plt.setp(subplot.xaxis.get_majorticklabels(), rotation=90)
ax1.set_xticklabels([])
ax2.set_xticklabels([])
ax3.set_xticklabels([])
ax1.set_ylim((0, 30))
ax5.set_ylim((0, 30))
ax1.set_ylabel('Average Euclidean Distance')
ax5.set_ylabel('Average Euclidean Distance')
ax7.legend(handles=LEGEND_ELEMENTS, bbox_to_anchor=(2.0, 0.4))
plt.show()