def squat_down(angle, angles_arr, squat_knee_angle, left_knee_angle, right_knee_angle, left_hip_gap, right_hip_gap):
global color
global test
global count_flag
global squat_count
global video_player#비디오 재생 트리거
# 내려갔을 때
x=""
if mean(angles_arr[-10:-5]) < mean(angles_arr[-5:]) and angle - angles_arr[0]>=10 :
if left_knee_angle > squat_knee_angle * 1.2 and right_knee_angle > squat_knee_angle * 1.2:
test="Lower"
color = (0,0,250)
elif (squat_knee_angle <= left_knee_angle < squat_knee_angle * 1.2 or
squat_knee_angle <= right_knee_angle < squat_knee_angle * 1.2) or \
(left_hip_gap < 20 or right_hip_gap < 20):
test="Good"
video_player=True #다 앉았을때 video재생 준비 완료
count_flag = True
color = (255,0,0)
x= "down"
# 올라올 때
if mean(angles_arr[-10:-5]) > mean(angles_arr[-5:]):
x="up"
# 내려갔다 올라와서 멈출 때
if min(angles_arr) * 0.9 < angle < min(angles_arr) * 1.1: #완전히 "섰다"의 인식이 쫌 여유가 있는 듯 합니다
x = 'ready'
if count_flag:
video_player=True#다 서있을 때 video 재생 준비 완료
count_flag = False
squat_count +=1
print(f"rep: {squat_count}")
def squat_down(angle, angles_arr, squat_knee_angle, left_knee_angle,
right_knee_angle, left_hip_gap, right_hip_gap, knee_gap):
global color
global test
global count_flag
global squat_count
global video_player #비디오 재생 트리거
global squat_set
global running
global rest_flag
global knee_flag
# 내려갔을 때
if len(angles_arr) > 10:
if mean(angles_arr[-10:-5]) < mean(
angles_arr[-5:]) and angle > angles_arr[0] + 5:
if left_knee_angle > squat_knee_angle * 1.2 and right_knee_angle > squat_knee_angle * 1.2:
test = "Lower"
color = (0, 0, 250)
elif (squat_knee_angle <= left_knee_angle < squat_knee_angle * 1.2 or
squat_knee_angle <= right_knee_angle < squat_knee_angle * 1.2) or \
(left_hip_gap < 20 or right_hip_gap < 20):
test = "Good"
video_player = 0 #다 앉았을때 video재생 준비 완료
count_flag = True
color = (255, 0, 0)
# 무릎이 너무 모이는 경우
if knee_gap >= 1.2:
knee_flag = True
# 올라올 때
if mean(angles_arr[-10:-5]) > mean(angles_arr[-5:]):
# 내려갔다 올라와서 멈출 때
if np.median(angles_arr[:5]) * 0.95 <= angle <= np.median(
angles_arr[:5]) * 1.05: #디테일 조정함@@@@@@@@@@@@@@@@@@@@@@@
if count_flag:
video_player = 1 #다 서있을 때 video 재생 준비 완료
print(angle)
count_flag = False
squat_count += 1
if squat_count - CUSTOM_SQUAT_COUNT == 0:
squat_set += 1
squat_count = 0
running = False
rest_flag = True
print(f"rep: {squat_count}")
# 무릎이 너무 모이는 경우
if knee_flag:
test = "Knee Wider"
color = (0, 0, 255)
knee_flag = False
def do_stuff():
rolls = np.array(return_med_index(x))
rolls = rolls.astype(int)
#print(arr)
arr.sort()
s0 = mean(arr)
s1 = mean(arr[rolls])
s2 = stats.mode(arr)[0].astype(float)[0]
print(round(s0, 1))
print(round(s1, 1))
print(round(s2, 1))
def filter_CS_candidates(self, vehicles, charging_stations):
d = great_circle_distance(vehicles.lat.values, vehicles.lon.values,
mean(charging_stations[:, 0]),
mean(charging_stations[:, 1]))
within_limit_distance = d < 1e3 * (
self.reject_distance + self.unit_length *
(self.k - 1)) # a large enough number
candidates = vehicles.index[within_limit_distance]
d = d[within_limit_distance]
return candidates[np.argsort(d)[:2 * len(charging_stations) +
1]].tolist()
def log_stats(ds_data: DsDataList, text_data: TextDataList):
stats: List[str, int, float, float, float] = []
text_lengths = [len(x.symbols) for x in text_data.items()]
stats.append((
"Overall",
len(text_lengths),
min(text_lengths),
max(text_lengths),
mean(text_lengths),
sum(text_lengths),
))
speakers_text_lengths: Dict[Speaker, List[float]] = {}
for ds_entry, text_entry in zip(ds_data.items(), text_data.items()):
if ds_entry.speaker_name not in speakers_text_lengths:
speakers_text_lengths[ds_entry.speaker_name] = []
speakers_text_lengths[ds_entry.speaker_name].append(
len(text_entry.symbols))
for speaker, speaker_text_lengths in speakers_text_lengths.items():
stats.append((
speaker,
len(speaker_text_lengths),
min(speaker_text_lengths),
max(speaker_text_lengths),
mean(speaker_text_lengths),
sum(speaker_text_lengths),
))
stats.sort(key=lambda x: (x[-1]), reverse=True)
stats_csv = pd.DataFrame(stats,
columns=[
"Speaker",
"# Entries",
"# Min",
"# Max",
"# Avg",
"# Total",
])
logger = getLogger(__name__)
with pd.option_context(
'display.max_rows',
None,
'display.max_columns',
None,
'display.width',
None,
'display.precision',
0,
):
logger.info(stats_csv)
def log_stats(ds_data: DsDataList, wav_data: WavDataList):
logger = getLogger(__name__)
if len(wav_data) > 0:
logger.info(f"Sampling rate: {wav_data.items()[0].wav_sampling_rate}")
stats: List[str, int, float, float, float, int] = []
durations = [x.wav_duration for x in wav_data.items()]
stats.append((
"Overall",
len(wav_data),
min(durations),
max(durations),
mean(durations),
sum(durations) / 60,
sum(durations) / 3600,
))
speaker_durations: Dict[Speaker, List[float]] = {}
for ds_entry, wav_entry in zip(ds_data.items(), wav_data.items()):
if ds_entry.speaker_name not in speaker_durations:
speaker_durations[ds_entry.speaker_name] = []
speaker_durations[ds_entry.speaker_name].append(wav_entry.wav_duration)
for speaker_name, speaker_durations in speaker_durations.items():
stats.append((
speaker_name,
len(speaker_durations),
min(speaker_durations),
max(speaker_durations),
mean(speaker_durations),
sum(speaker_durations) / 60,
sum(speaker_durations) / 3600,
))
stats.sort(key=lambda x: (x[-2]), reverse=True)
stats_csv = pd.DataFrame(stats, columns=[
"Speaker",
"# Entries",
"Min (s)",
"Max (s)",
"Avg (s)",
"Total (min)",
"Total (h)",
])
with pd.option_context(
'display.max_rows', None,
'display.max_columns', None,
'display.width', None,
'display.precision', 4,
):
print(stats_csv)
def __get_ordering(self, obs1: List[float], obs2: List[float],
p_value: float) -> Literal['<', '=', '>']:
"""
Compares the p-value from the t-test of the two samples and returns the statistical significant ordering
The '=' ordering operator is used to signify no statistical significant ordering for the two samples
"""
P_LIMIT = 0.05
if p_value < P_LIMIT:
diff = mean(obs1) - mean(obs2)
if diff > 0:
return '>'
elif diff < 0:
return '<'
return '='
def main():
get_args()
def sentences():
return chain.from_iterable(
(read_slice(data) for data in read_corpus()))
bigram = Phrases(sentences(), min_count=1, threshold=1, delimiter=b' ')
bigram_phraser = Phraser(bigram)
bigrammed = map(lambda x: bigram_phraser[x], sentences())
trigram = Phrases(bigrammed, min_count=1, threshold=1, delimiter=b' ')
trigram_phraser = Phraser(trigram)
only_trigrams = {b' '.join(trigram_tuple): score for (trigram_tuple, score) in \
trigram_phraser.phrasegrams.items() if b' '.join(trigram_tuple).count(b' ') == 2}
for key, value in sorted(only_trigrams.items(),
key=lambda item: item[1],
reverse=True)[:10]:
print(key, value)
scores = list(only_trigrams.values())
print("""
Unique trigrams: {unique}
Mean score:{mean}
Max score:{max}
Min score:{min}
""".format(unique=len(only_trigrams),
mean=mean(scores) if len(scores) != 0 else 0,
max=max(scores) if len(scores) != 0 else 0,
min=min(scores) if len(scores) != 0 else 0))
def plotTable(v, vol, prob, lbds, MOQ, name):
"""
DEFINITION:
v: index in volatility array
vol: volatility array. Contains simulated volatilities.
name: name of the plot.
"""
np, _, nr, _ = shape(MOQ)
MOQshape = zeros((np, nr))
for p in xrange(len(prob)):
for r in xrange(len(lbds)):
MOQshape[p, r] = mean(MOQ[p, v, r, :])
fig = plt.figure()
ax = fig.gca(projection='3d')
X = lbds
Y = prob
X, Y = meshgrid(X, Y)
Z = MOQshape
plt.xlabel('Lambdas')
plt.ylabel('Probabilities')
plt.title(name + ' for fixed lambdas - v=' + str(vol[v]))
surf = ax.plot_surface(X,
Y,
Z,
rstride=1,
cstride=1,
cmap=cm.jet,
linewidth=0,
antialiased=False)
ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
fig.colorbar(surf, shrink=0.5, aspect=5)
def _collapse_column(self, column_name, current_value):
"""Here is where values are collapsed based on the input values. Assumed t be separated by a "|"
current_value is assumed to be a string.
"""
if column_name not in self._columns_to_collapse:
return current_value
list_vals = current_value.split("|")
no_blank_vals = [v for v in list_vals if v.strip() != "" and v.strip() != "."]
# if no "." is found in the value, assume that this should be rendered as an int, if MIN was chosen
# we assume it is an int if no "." is found.
is_assuming_int = all([v.find(".") == -1 for v in no_blank_vals])
try:
if is_assuming_int:
final_vals = [int(v) for v in no_blank_vals]
else:
final_vals = [float(v) for v in no_blank_vals]
except ValueError as te:
logging.getLogger(__name__).warn("Could not collapse " + column_name + ":" + current_value + " into one number. Returning the input value.")
return current_value
if len(final_vals) == 0:
return ""
if self._method_dict[column_name] == ColumnCollapser.MIN:
return str(min(final_vals))
elif self._method_dict[column_name] == ColumnCollapser.MEAN:
return str(mean(final_vals))
def PlotWss(self, meshid, imagpath):
'''
This method plots Wss signal and returns peak wss.
'''
try:
import matplotlib
matplotlib.use('Agg') #switch to matplotlib.use('WXAgg') if you want to show and not save velocity profile.
from matplotlib.pyplot import plot, xlabel, ylabel, title, legend, savefig, close, ylim
except:
sys.exit("PlotWss method requires matplotlib package (http://matplotlib.sourceforge.net.\n")
tplot = linspace(0, self.tPeriod, len(self.Tauplot))
plot(tplot, self.Tauplot,'g-',linewidth = 3, label = 'WSS')
minY = 0
for w in self.Tauplot:
if w < minY:
minY = w
if minY != 0:
plot(tplot, zeros(len(self.Tauplot)),':',linewidth = 1)
ylim(ymin=minY)
xlabel('Time ($s$)')
ylabel('Wall shear stress ($dyne/cm^2$)')
title ('Wss'+' peak:'+str(round(max(self.Tauplot),1))+' mean:'+str(round(mean(self.Tauplot),1))+' min:'+str(round(min(self.Tauplot),1)))
legend()
savefig(imagpath+str(meshid)+'_'+str(self.Name)+'_wss.png')
print "Wss, MeshId", meshid, self.Name, "=", str(round(max(self.Tauplot),1)), "$dyne/cm^2$"
close()
return (round(max(self.Tauplot),1))
def trans_lcqmc_bert(dataset: list, vocab: Vocabulary, is_merge=0):
"""
最大长度
"""
out_arr, text_len = [], []
for each in dataset:
t1, t2, label = each.text_a, each.text_b, int(each.label)
if is_merge:
out_ids1, mask_ids1, seg_ids1, seq_len1 = vocab._transform_2seq2bert_id(
t1, t2, padding=1)
out_arr.append([out_ids1, mask_ids1, seg_ids1, seq_len1, label])
text_len.extend([len(t1) + len(t2)])
else:
out_ids1, mask_ids1, seg_ids1, seq_len1 = vocab._transform_seq2bert_id(
t1, padding=1)
out_ids2, mask_ids2, seg_ids2, seq_len2 = vocab._transform_seq2bert_id(
t2, padding=1)
out_arr.append([
out_ids1, mask_ids1, seg_ids1, seq_len1, out_ids2, mask_ids2,
seg_ids2, seq_len2, label
])
text_len.extend([len(t1), len(t2)])
pass
print("max len", max(text_len), "avg len", mean(text_len), "cover rate:",
np.mean([x <= conf.max_seq_len for x in text_len]))
return out_arr
def normalizeData(data, meanOnly = False):
"""
normalize data by subtracting mean and dividing by sd per COLUMN
@param data: an array
@param meanOnly: if True subtract mean only; otherwise divide by sd too
@return: (an array with the same dimension as data, mean, stds, the transformer)
"""
# compute the new data
m = mean(data, axis=0)
res = data - m
if meanOnly:
stds = 1
else:
stds = sqrt(var(data, axis=0))
stds[stds==0] = 1 # to avoid dividing by 0
res /= stds
# figure out the transformer
def foo(givenData):
assert givenData.shape[1]==data.shape[1], "Only arrays of %d columns are handled." % data.shape[1]
return (givenData - m)/stds
return res, m, stds, foo
def printMe(self):
"""print the result of doVse in an accessible format.
for instance:
vses.printMe()
"""
for i in range(len(self.methods)):
print(self.methods[i][0])
print([strat.getName() for strat in self.methods[i][1]],
[mean([result[i].results[j].result[0] for result in self.vses])
for j in range(len(self.methods[i][1]) - 1)],
mean(
[(0 if result[i].results[0].result[0]==result[i].results[2].result[0] else 1)
for result in self.vses]
)
)
def normalizeData(data, meanOnly = False):
"""
normalize data by subtracting mean and dividing by sd per COLUMN
Parameters:
- data: an array
- meanOnly: if True subtract mean only; otherwise divide by sd too
Returns: an array with the same dimension as data
"""
if meanOnly:
return data-mean(data,axis=0)
else:
stds = sqrt(var(data, axis=0))
stds[stds==0] = 1 # to avoid dividing by 0
return (data - mean(data, axis=0))/stds
def _phase_3(self, npoints):
normalized_edges_dict = {}
to_remove = set()
for node in range(npoints):
# transforms int, int, dict into int, int, float (src, dst, weight)
tmp = set(
Edge.tuple_representation(Edge.convert_weight_dict(t))
for t in self.g.edges(node, data=True))
normalized_edges_dict[node] = tmp
for node in range(npoints):
local_edges = normalized_edges_dict[node].copy()
edges2 = local_edges
for neigh in local_edges:
other = neigh[1]
other_edges = normalized_edges_dict[other]
edges2 = edges2.union(other_edges)
mean2 = mean(list(map(lambda e: e[-1],
edges2))) if len(edges2) > 0 else 0
self.local_mean2.append(mean2)
for e in edges2:
w = e[-1]
if w > (mean2 + self.mean_std_dev):
to_remove.add((e[0], e[1]))
self.local_mean2 = np.array(self.local_mean2)
for e in to_remove:
self.g.remove_edge(e[0], e[1])
self.labels_, self.cluster_sizes = self._label_conn_comp(
self.shuffle_labels)
return
def GetTaoFromQ(self, el):
'''
Computing wall shear stress in terms of the flow rate,
using inverse womersley method of Cezeaux et al.1997
'''
self.radius = mean(el.Radius)
self.Res = el.R
self.length = el.Length
self.Name = el.Name
#WOMERSLEY NUMBER
self.alpha = self.radius * sqrt(
(2.0 * pi * self.density) / (self.tPeriod * self.viscosity))
#FOURIER SIGNAL
k = len(self.signal)
n = 0
while n < (self.nHarmonics):
An = 0
Bn = 0
for i in arange(k):
An += self.signal[i] * cos(
n * (2.0 * pi / self.tPeriod) * self.dt * self.nSteps[i])
Bn += self.signal[i] * sin(
n * (2.0 * pi / self.tPeriod) * self.dt * self.nSteps[i])
An = An * (2.0 / k)
Bn = Bn * (2.0 / k)
self.fourierModes.append(complex(An, Bn))
n += 1
self.Steps = linspace(0, self.tPeriod, self.samples)
self.WssSignal = []
self.Tauplot = []
for step in self.Steps:
self.tao = -self.fourierModes[0].real * 2.0
k = 1
while k < self.nHarmonics:
cI = complex(0., 1.)
cA = (self.alpha * pow((1.0 * k), 0.5)) * pow(cI, 1.5)
c1 = 2.0 * jn(1, cA)
c0 = cA * jn(0, cA)
cT = complex(0, -2.0 * pi * k * self.t / self.tPeriod)
'''tao computation'''
taoNum = self.alpha**2 * cI**3 * jn(1, cA)
taoDen = c0 - c1
taoFract = taoNum / taoDen
cTao = self.fourierModes[k] * exp(cT) * taoFract
self.tao += cTao.real
k += 1
self.tao *= -(self.viscosity / (self.radius**3 * pi))
self.Tauplot.append(self.tao * 10) #dynes/cm2
self.WssSignal.append(self.tao)
self.t += self.dtPlot
return self.WssSignal #Pascal
def _mean_vote(self, all_predictions):
"""Returns mean of the predictions
Args:
all_predictions (List[List[float]]): The predictions from all models, per event
Returns:
List[Float]: The mean of predictions from all models, per event
"""
return [mean(predictions) for predictions in all_predictions]
def getSensitivity(self):
allSens = []
for example in self.examples:
sens = []
modifiedExample = example
isMyClass = -1.0
if example["digit"] == self.myDigit:
if example["isProper"]:
isMyClass = 1.0
result = self.classify(example["pixelMap"])
for i in range(len(example["pixelMap"])):
modifiedExample["pixelMap"][
i] = modifiedExample["pixelMap"][i] + self.H
modifiedResult = self.classify(modifiedExample["pixelMap"])
sens.append((modifiedResult - result) / self.H)
allSens.append(mean(sens))
self.sensitivity = mean(allSens)
print("Sensitivity for digit %s: %s" %
(self.myDigit, self.sensitivity))
def GetTaoFromQ(self,el):
'''
Computing wall shear stress in terms of the flow rate,
using inverse womersley method of Cezeaux et al.1997
'''
self.radius = mean(el.Radius)
self.Res = el.R
self.length = el.Length
self.Name = el.Name
#WOMERSLEY NUMBER
self.alpha = self.radius * sqrt((2.0 *pi*self.density)/(self.tPeriod*self.viscosity))
#FOURIER SIGNAL
k = len(self.signal)
n = 0
while n < (self.nHarmonics):
An = 0
Bn = 0
for i in arange(k):
An += self.signal[i] * cos(n*(2.0*pi/self.tPeriod)*self.dt*self.nSteps[i])
Bn += self.signal[i] * sin(n*(2.0*pi/self.tPeriod)*self.dt*self.nSteps[i])
An = An * (2.0/k)
Bn = Bn * (2.0/k)
self.fourierModes.append(complex(An, Bn))
n+=1
self.Steps = linspace(0,self.tPeriod,self.samples)
self.WssSignal = []
self.Tauplot = []
for step in self.Steps:
self.tao = -self.fourierModes[0].real * 2.0
k=1
while k < self.nHarmonics:
cI = complex(0.,1.)
cA = (self.alpha * pow((1.0*k),0.5)) * pow(cI,1.5)
c1 = 2.0 * jn(1, cA)
c0 = cA * jn(0, cA)
cT = complex(0, -2.0*pi*k*self.t/self.tPeriod)
'''tao computation'''
taoNum = self.alpha**2*cI**3*jn(1,cA)
taoDen = c0-c1
taoFract = taoNum/taoDen
cTao = self.fourierModes[k] * exp(cT) * taoFract
self.tao += cTao.real
k+=1
self.tao *= -(self.viscosity/(self.radius**3*pi))
self.Tauplot.append(self.tao*10) #dynes/cm2
self.WssSignal.append(self.tao)
self.t += self.dtPlot
return self.WssSignal #Pascal
def getMostSevereValue(self, minNInstants=1): # TODO use np.percentile
from matplotlib.mlab import find
from numpy.core.multiarray import array
from numpy.core.fromnumeric import mean
values = array(self.values.values())
indices = range(len(values))
if len(indices) >= minNInstants:
values = sorted(values[indices], reverse = self.mostSevereIsMax) # inverted if most severe is max -> take the first values
return mean(values[:minNInstants])
else:
return None
def main():
os.chdir(r'G:\Shared drives\Apex\Acoustic Data\IOA, conference proceedings\2021 papers\Max noise levels from hockey pitches')
# adjust_value_by_LAeqT()
LAFmax = []
filenames = [str(1+n)+'.wav' for n in range(104)]
LAFmax = [calc_LAeq_dt(os.path.join('bat hitting', fn)) for fn in filenames]
hist_plot(LAFmax)
print(mean(LAFmax))
print(std(LAFmax))
plt.xlabel('LAFmax at 11 m')
plt.show()
def squat_down(angle, angles_arr, squat_knee_angle, left_knee_angle, right_knee_angle, left_hip_gap, right_hip_gap):
global color
global test
global count_flag
global squat_count
global video_player#비디오 재생 트리거
global squat_set
global running
global rest_flag
# 내려갔을 때
if mean(angles_arr[-10:-5]) < mean(angles_arr[-5:]) and angle - angles_arr[0]>=10 :
if left_knee_angle > squat_knee_angle * 1.2 and right_knee_angle > squat_knee_angle * 1.2:
test="Lower"
color = (0,0,250)
elif (squat_knee_angle <= left_knee_angle < squat_knee_angle * 1.2 or
squat_knee_angle <= right_knee_angle < squat_knee_angle * 1.2) or \
(left_hip_gap < 20 or right_hip_gap < 20):
test="Good"
video_player=0 #다 앉았을때 video재생 준비 완료
count_flag = True
color = (255,0,0)
# 올라올 때
if mean(angles_arr[-10:-5]) > mean(angles_arr[-5:]):
# 내려갔다 올라와서 멈출 때
if min(angles_arr) * 0.9 < angle < min(angles_arr) * 1.1: #완전히 "섰다"의 인식이 쫌 여유가 있는 듯 합니다
if count_flag:
video_player=1#다 서있을 때 video 재생 준비 완료
count_flag = False
squat_count +=1
if squat_count % CUSTOM_SQUAT_COUNT == 0:
squat_set += 1
squat_count = 0
running = False
rest_flag = True
print(f"rep: {squat_count}")
def GetRadius(self, abscissa):
'''
This method returns edge's radius
'''
if 'value' in self.Radius:
return self.Radius['value']
if 'array' in self.Radius:
if abscissa is not None:
return self.Radius['array'][abscissa]
else:
if self.edgeAbscissa is None:
return mean(array((self.Radius['array'].values())))
else:
return self.Radius['array'][self.edgeAbscissa]
def Classic(inputImage, decodedObjects, X0):
Data = []
Distance = []
dY = []
for i in range(0, len(decodedObjects)):
zbarData = decodedObjects[i].data
arr = list(map(float, zbarData.split()))
Data.append([arr[2], arr[3]])
polygon = decodedObjects[i].polygon
SIDE_OF_QR = arr[0]
H_QR = arr[1] - H_CAMERA
data = polygon[:]
if ((polygon[0].y + 30) < (polygon[1].y)):
data = polygon[:]
key = False
else:
key = True
data[0] = polygon[3]
data[1] = polygon[0]
data[2] = polygon[1]
data[3] = polygon[2]
centerTop = getCenter(data[0], data[3])
centerBottom = getCenter(data[1], data[2])
a = distanceCalculate2(data[0], data[1], H_QR, SIDE_OF_QR)
b = distanceCalculate2(centerTop, centerBottom, H_QR, SIDE_OF_QR)
d = distanceCalculate2(data[2], data[3], H_QR, SIDE_OF_QR)
b = mean([a, b, d])
#dY.append( coordY(centerTop, centerBottom,(centerTop.x+centerBottom.x)/2.,SIDE_OF_QR) ) альтернативный метод
dy = coordY(centerTop, centerBottom,
(centerTop.x + centerBottom.x) / 2., SIDE_OF_QR)
b = (b * b + dy * dy)
Distance.append(b)
res = scipy.optimize.leastsq(RP, X0, args=(Distance, Data))
x = res[0]
#dy=(dY[0]+dY[1])/2 алтернативный метод
#x[1]=x[1]-dy
cv2.putText(inputImage,
f"X(gl) = {round(x[0], 3)}, Y(gl) = {round(x[1],3)} ",
(320, 110), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 0, 255), 2,
cv2.LINE_AA)
return x[0], x[1]
def get_rewards(file):
rewards = []
steps = []
with open(file) as f:
for l in f.readlines()[2:]:
rewards.append(float(l.split(",")[0]))
steps.append(int(l.split(",")[1]))
steps = list(np.cumsum(steps))
final_r = []
final_s = []
for i in range(1000, len(rewards), 1):
final_r.append(mean(rewards[i - 1000:i]))
final_s.append(steps[i])
return final_r, final_s
def GetRadius(self, abscissa):
'''
This method returns edge's radius
'''
if 'value' in self.Radius:
return self.Radius['value']
if 'array' in self.Radius:
if abscissa is not None:
return self.Radius['array'][abscissa]
else:
if self.edgeAbscissa is None:
return mean(array((self.Radius['array'].values())))
else:
return self.Radius['array'][self.edgeAbscissa]
def calc_LAeq0800_1800(self):
all_data = []
for n in range(self.days):
day_n = self.start_date + pd.Timedelta(n, "days")
start_time = day_n + pd.Timedelta(8, "hours")
end_time = day_n + pd.Timedelta(18, "hours")
df = self.noise[(self.noise['Time_obj'] >= start_time)
& (self.noise['Time_obj'] < end_time)]
LAeq0800_1800 = 10. * np.log10(mean(10**(df['Leq'].values / 10)))
all_data.append(LAeq0800_1800)
output = pd.DataFrame(all_data,
index=self.day_all,
columns=['LAeq0800_1800'])
output = output.transpose()
output.to_csv('LAeq0800_1800.csv')
def vseOn(self, voters, chooserFuns=(), **args):
"""Finds honest and strategic voter satisfaction efficiency (VSE)
for this method on the given electorate.
"""
multiResults = self.multiResults(voters, chooserFuns, **args)
utils = voters.socUtils
best = max(utils)
rand = mean(utils)
#import pprint
#pprint.pprint(multiResults)
vses = VseMethodRun(self.__class__, chooserFuns,
[VseOneRun([(utils[self.winner(result)] - rand) / (best - rand)],tally,chooser)
for ((result, chooser), tally) in multiResults[0]])
vses.extraEvents=multiResults[1]
return vses
def vseOn(self, voters, chooserFuns=(), **args):
"""Finds honest and strategic voter satisfaction efficiency (VSE)
for this method on the given electorate.
"""
multiResults = self.multiResults(voters, chooserFuns, **args)
utils = voters.socUtils
best = max(utils)
rand = mean(utils)
#import pprint
#pprint.pprint(multiResults)
vses = VseMethodRun(self.__class__, chooserFuns,
[VseOneRun([(utils[self.winner(result)] - rand) / (best - rand)],tally,chooser)
for (result, chooser, tally) in multiResults[0]])
vses.extraEvents=multiResults[1]
return vses
def calculateFitness(self, population, _):
if not self.inited:
self.inited = True
return
sys.stdout.flush()
all = []
for list in population:
for ind in list["individuals"]:
all.append(ind)
num_threads = int(mp.cpu_count() - 1)
pool = mp.Pool(num_threads)
res = pool.map(self.battle, all)
for ind, fitness in zip(all, res):
ind.setFitness(fitness)
average = mean(res)
print(f"average fitness {average}")
pool.close()
def trans_lcqmc(dataset):
"""
最大长度
"""
out_arr, text_len = [], []
for each in dataset:
t1, t2, label = each.text_a, each.text_b, int(each.label)
t1_ids = convert_word2id(t1, conf.vocab_map)
t1_len = conf.max_seq_len if len(t1) > conf.max_seq_len else len(t1)
t2_ids = convert_word2id(t2, conf.vocab_map)
t2_len = conf.max_seq_len if len(t2) > conf.max_seq_len else len(t2)
# t2_len = len(t2)
out_arr.append([t1_ids, t1_len, t2_ids, t2_len, label])
# out_arr.append([t1_ids, t1_len, t2_ids, t2_len, label, t1, t2])
text_len.extend([len(t1), len(t2)])
pass
print("max len", max(text_len), "avg len", mean(text_len), "cover rate:", np.mean([x <= conf.max_seq_len for x in text_len]))
return out_arr
def evaluate_populations(swarm, envs: List, eval_episode: int) -> Tuple[float, float]:
scores = np.zeros(shape=len(envs))
for i, agent in enumerate(swarm.population):
env = envs[i]
score = 0
for i_episode in range(eval_episode):
obs = env.reset()
while True:
action = agent.choose_action(obs, use_noise=False)
next_obs, reward, done, _ = env.step(action)
score += reward
if done:
break
else:
obs = next_obs
scores[i] = score / eval_episode
return mean(scores), max(scores)
def clean_raw_data(df):
""" Takes a dataframe and performs four steps:
- Selects columns for modeling
- For numeric variables, replaces 0 values with mean for that region
- Fills invalid construction_year values with the mean construction_year
- Converts strings to categorical variables
:param df: A raw dataframe that has been read into pandas
:returns: A dataframe with the preprocessing performed.
"""
useful_columns = [
'amount_tsh', 'gps_height', 'longitude', 'latitude', 'region',
'population', 'construction_year', 'extraction_type_class',
'management_group', 'quality_group', 'source_type', 'waterpoint_type',
'status_group'
]
df_input = df[useful_columns].copy()
zero_is_bad_value = ['longitude', 'population']
other_bad_value = ['latitude']
# df_input = replace_value_with_grouped_mean(df_input)
# df_input = replace_value_with_grouped_mean(df_input, np.nan, 'construction_mean')
for col in useful_columns:
if df_input[col].dtype == 'object':
df_input[col] = df_input[col].astype("category")
# print('change col {} format'.format(col))
if col == 'construction_year':
invalid_rows = df_input[col] <= 1000
valid_mean = mean(df_input.loc[(~invalid_rows), col])
df_input.loc[invalid_rows, col] = valid_mean
# print('change all construction year less than 1000 to mean')
if col in zero_is_bad_value:
df_input = replace_value_with_grouped_mean(df_input, 0, col,
'region')
print("Change col {} from 0 to mean".format(col))
if col in other_bad_value:
df_input = replace_value_with_grouped_mean(df_input, -2e-8, col,
'region')
print("Change col {} from -2e-8 to mean".format(col))
return df_input
def indicatorMap(indicatorValues, trajectory, squareSize):
'''Returns a dictionary
with keys for the indices of the cells (squares)
in which the trajectory positions are located
at which the indicator values are attached
ex: speeds and trajectory'''
from numpy import floor, mean
assert len(indicatorValues) == trajectory.length()
indicatorMap = {}
for k in xrange(trajectory.length()):
p = trajectory[k]
i = floor(p.x/squareSize)
j = floor(p.y/squareSize)
if indicatorMap.has_key((i,j)):
indicatorMap[(i,j)].append(indicatorValues[k])
else:
indicatorMap[(i,j)] = [indicatorValues[k]]
for k in indicatorMap.keys():
indicatorMap[k] = mean(indicatorMap[k])
return indicatorMap
def resultsTable(self, eid, emodel, cands, voters, chooserFuns=(), **args):
multiResults = self.multiResults(voters, chooserFuns, **args)
utils = voters.socUtils
best = max(utils)
rand = mean(utils)
rows = list()
nvot=len(voters)
for ((result, chooser), tally) in multiResults[0]:
row = {
"eid":eid,
"emodel":emodel,
"ncand":cands,
"nvot":nvot,
"best":best,
"rand":rand,
"method":str(self),
"chooser":chooser,#.getName(),
"util":utils[self.winner(result)],
"vse":(utils[self.winner(result)] - rand) / (best - rand)
}
#print(tally)
for (i, (k, v)) in enumerate(tally.items()):
#print("Result: tally ",i,k,v)
row["tallyName"+str(i)] = str(k)
row["tallyVal"+str(i)] = str(v)
rows.append(row)
if len(multiResults[1]):
row = {
"eid":eid,
"emodel":emodel,
"method":self.__class__.__name__,
"chooser":"extraEvents",
"util":None
}
for (i, (k, v)) in enumerate(multiResults[1]):
row["tallyName"+str(i)] = str(k)
row["tallyVal"+str(i)] = str(v)
rows.append(row)
return(rows)
def resultsTable(self, eid, emodel, cands, voters, chooserFuns=(), **args):
multiResults = self.multiResults(voters, chooserFuns, **args)
utils = voters.socUtils
best = max(utils)
rand = mean(utils)
rows = list()
nvot = len(voters)
for (result, chooser, tallyItems) in multiResults:
row = {
"eid": eid,
"emodel": emodel,
"ncand": cands,
"nvot": nvot,
"best": best,
"rand": rand,
"method": str(self),
"chooser": chooser, #.getName(),
"util": utils[self.winner(result)],
"vse": (utils[self.winner(result)] - rand) / (best - rand)
}
#print(tallyItems)
for (i, (k, v)) in enumerate(tallyItems):
#print("Result: tally ",i,k,v)
row["tallyName" + str(i)] = str(k)
row["tallyVal" + str(i)] = str(v)
rows.append(row)
# if len(multiResults[1]):
# row = {
# "eid":eid,
# "emodel":emodel,
# "method":self.__class__.__name__,
# "chooser":"extraEvents",
# "util":None
# }
# for (i, (k, v)) in enumerate(multiResults[1]):
# row["tallyName"+str(i)] = str(k)
# row["tallyVal"+str(i)] = str(v)
# rows.append(row)
return (rows)
def PlotWss(self, meshid, imagpath):
'''
This method plots Wss signal and returns peak wss.
'''
try:
import matplotlib
matplotlib.use(
'Agg'
) #switch to matplotlib.use('WXAgg') if you want to show and not save velocity profile.
from matplotlib.pyplot import plot, xlabel, ylabel, title, legend, savefig, close, ylim
except:
sys.exit(
"PlotWss method requires matplotlib package (http://matplotlib.sourceforge.net.\n"
)
tplot = linspace(0, self.tPeriod, len(self.Tauplot))
plot(tplot, self.Tauplot, 'g-', linewidth=3, label='WSS')
minY = 0
for w in self.Tauplot:
if w < minY:
minY = w
if minY != 0:
plot(tplot, zeros(len(self.Tauplot)), ':', linewidth=1)
ylim(ymin=minY)
xlabel('Time ($s$)')
ylabel('Wall shear stress ($dyne/cm^2$)')
title('Wss' + ' peak:' + str(round(max(self.Tauplot), 1)) + ' mean:' +
str(round(mean(self.Tauplot), 1)) + ' min:' +
str(round(min(self.Tauplot), 1)))
legend()
savefig(imagpath + str(meshid) + '_' + str(self.Name) + '_wss.png')
print "Wss, MeshId", meshid, self.Name, "=", str(
round(max(self.Tauplot), 1)), "$dyne/cm^2$"
close()
return (round(max(self.Tauplot), 1))
def problem2(data, figureDir, dataName, l, maxNumRepetitions, minSampleSize, targetValue):
if os.path.exists(getProblem2FigureLoc(figureDir, dataName, l, maxNumRepetitions)):
#this implies that all figures before it have been created already, so we don't need to repeat them
return
MSEValues = [[] for x in xrange(minSampleSize, len(data[TRAIN]) + 1)]
sampleSizeValueList = range(minSampleSize, len(data[TRAIN]) + 1, 1)
sampleSizeValueArray = array(sampleSizeValueList)
targetArray = targetValue * ones(len(sampleSizeValueList), dtype=numpy.float64)
for repeatNum in range(1, maxNumRepetitions+1):
#randomly choose ordering of the samples for this run
#make the range of indexes, then shuffle them into a random order
randomlySortedIndexes = range(len(data[TRAIN]))
shuffle(randomlySortedIndexes)
#start with a sample size of one, go to the total training set
for sampleSizeIndex, sampleSize in enumerate(sampleSizeValueList):
curSampleIndexesList = randomlySortedIndexes[:sampleSize]
curTrainSample = selectSample(data[TRAIN], curSampleIndexesList)
curTrainLabelSample = selectSample(data[TRAIN_LABELS], curSampleIndexesList)
w = doubleU(phi(curTrainSample), l, tListToTVector(curTrainLabelSample))
curSampleMSE = MSE(data[TEST], w, data[TEST_LABELS])
MSEValues[sampleSizeIndex].append(squeeze(curSampleMSE))
#have a sample size of 0 is meaningless
curRepeatMeanMSEValues = array([mean(array(x, dtype=numpy.float64)) for x in MSEValues])
plt.plot(sampleSizeValueArray, curRepeatMeanMSEValues, '-', label="Learning curve")
plt.plot(sampleSizeValueArray, targetArray, '--', label="Target MSE")
plt.title("lamba = " + str(l) + " - " + str(repeatNum) + " repetitions")
plt.xlabel("Sample Size - minimum " + str(minSampleSize))
plt.ylabel("MSE on Full Test Set")
plt.xlim(xmin=targetValue - .5)
plt.legend(loc=0)
plt.savefig(getProblem2FigureLoc(figureDir, dataName, l, repeatNum))
plt.clf()
outputPath = path + 'statistics/'
begin = time.time()
prob = array([.55, .65, .75, .85, .95])
vol = array([0.0,0.001,0.005,0.01,0.05,0.1])
lRates = [0.01,0.11,0.21,0.31,0.41,0.51, 0.61, 0.71, 0.81, 0.91]
nEpisodes = 50
meanSquareError = np.zeros((len(prob), len(vol),len(lRates), nEpisodes))
rightEstimate = np.zeros((len(prob), len(vol), len(lRates),nEpisodes))
rightPrediction = np.zeros((len(prob), len(vol), len(lRates),nEpisodes))
rewardedTrials = np.zeros((len(prob), len(vol), len(lRates),nEpisodes))
totalIter = len(prob)*len(vol)*nEpisodes*len(lRates)
n=1 #iterations counter
for v in xrange(len(vol)):
for p in xrange(len(prob)):
environment = loadArrangeVar(prob[p], vol[v], path,'environment')
for r in xrange(len(lRates)):
for e in xrange(nEpisodes):
agent = loadCtLbdEpisodeVar(prob[p], vol[v], lRates[r],e, path,'agent')
meanSquareError[p][v][r][e] = mean(agent.err**2)
rightEstimate[p][v][r][e] = np.sum(around(agent.x[1:]) == around(environment.history)) / float(environment.history.size)*100 #x has a shape of nTrials+1 and history of nTrials. That is because after the last trial the agent learns the value of x for the next
rightPrediction[p][v][r][e] = np.sum(around(agent.x[0:-1]) == around(environment.history)) / float(environment.history.size)*100
rewardedTrials[p][v][r][e] = float(np.sum(agent.r))/agent.x.size*100 #Calculates how often the agent was rewarded within the episode
showProgress(totalIter, n, time.time(), begin)
n+=1
variables = {'meanSquareError':meanSquareError, 'rightEstimate':rightEstimate, 'rewardedTrials': rewardedTrials, 'rightPrediction':rightPrediction}
saveAllVars(outputPath,variables)
print 'Calculation finished in ', (time.time()-begin), 'seconds.'
def get_LA90():
#a weigting from 31.5 to 8k Hz in 1/3 octave
A_weighting = np.array([
-39.4, -34.6, -30.2, -26.2, -22.5, -19.1, -16.1, -13.4, -10.9, -8.6,
-6.6, -4.8, -3.2, -1.9, -0.8, 0, 0.6, 1, 1.2, 1.3, 1.2, 1, 0.5, -0.1,
-1.1
])
flower, fcentre, fupper = third_octave_bands(
) # fcentre from 31.5 Hz to 16k Hz
samplerate, data = wavfile.read('MS_PCM_signed_16bit.wav')
how_many_seconds = int(data.shape[0] / samplerate)
level_a_each_sec = []
adjust_list = []
LAeq1_4s = [47.3, 48.1, 48.9, 49.4]
# caluclate adjust values
for n in range(4):
data_slice = data[n * samplerate:(n + 1) * samplerate]
xf, levels = fft_wave_data(samplerate, data_slice, NFFT=16384 * 2)
fft_df = pd.DataFrame({'Frequency': xf, 'Level_dB': levels})
level_lin = []
for m in range(len(fcentre)):
low, fc, up = flower[m], fcentre[m], fupper[m]
temp_df = fft_df[(fft_df['Frequency'] >= low)
& (fft_df['Frequency'] < up)]
levels = temp_df['Level_dB'].values
Lp = 10. * np.log10(sum(10**(levels / 10)))
level_lin.append(Lp)
level_a_not_adjusted = np.array(level_lin) + A_weighting
total_a = 10. * np.log10(sum(10**(level_a_not_adjusted / 10)))
adjust_list.append(total_a - LAeq1_4s[n])
adjust = mean(adjust_list)
# calculuate the spectrum for each sec
for n in range(how_many_seconds):
# if n >100:
# break
data_slice = data[n * samplerate:(n + 1) * samplerate]
xf, levels = fft_wave_data(samplerate, data_slice, NFFT=16384 * 2)
fft_df = pd.DataFrame({'Frequency': xf, 'Level_dB': levels})
level_lin = []
for m in range(len(fcentre)):
low, fc, up = flower[m], fcentre[m], fupper[m]
temp_df = fft_df[(fft_df['Frequency'] >= low)
& (fft_df['Frequency'] < up)]
levels = temp_df['Level_dB'].values
Lp = 10. * np.log10(sum(10**(levels / 10)))
level_lin.append(Lp)
level_a = np.array(level_lin) + A_weighting - adjust
level_a_each_sec.append(level_a)
# write to pandas DataFrame and export to csv
freqs = ['%d' % int(c) for c in fcentre]
df = pd.DataFrame(np.array(level_a_each_sec), columns=freqs)
df.to_csv('LAeqT_per_sec.csv') # taking long time to write file !!
print(df.tail())
La90 = []
for f in freqs:
ss = df[f].sort_values() # ascending
La90.append(ss[int(0.1 * ss.shape[0])])
La90_df = pd.DataFrame({'Frequency': freqs, 'LA90_dB': La90})
La90_df.to_csv('La90_at_tird_octave.csv')
print(La90_df)
def _moving_mean(self):
"""
Returns mean of the last n prices
"""
return mean(self.last_n_prices)
def svr(C, gamma, eps):
#initialization of data wmproxy
traininput, traintarget, testinput, testtarget = initialize_wmproxy()
#training of the SVR
#scaling values in training and test targets
for i in range(len(traintarget)):
if(traintarget[i] != 0):
traintarget[i] = log(traintarget[i])
if(traininput[i] != 0):
traininput[i] = log(traininput[i])
for i in range(len(testtarget)):
if(testtarget[i] != 0):
testtarget[i] = log(testtarget[i])
if(testinput[i] != 0):
testinput[i] = log(testinput[i])
avg = mean(traintarget)
sigma = std(traintarget)
maxtrain = len(traintarget)
C = max([abs(avg + sigma), abs(avg - sigma)])
print "C is equal to %f" % C
svr = SVR(traininput[maxtrain-1440:maxtrain], testinput, traintarget[maxtrain-1440:maxtrain],gamma,C,eps,eps)
out = svr.svr_req(testinput[0:30])
error = 0
for i in range(len(out)):
error += (out[i] - testtarget[i])
mean_error = error / len(out)
variance = 0
for i in range(len(out)):
variance = abs(out[i] - mean_error)
variance /= len(out)
print "Variance = %f" % variance
epsilon = 3*variance*sqrt(log(len(out))/len(out))
print "Epsilon = %f" % epsilon
#calculation of the metrics
sme = svr.calc_sme(testtarget[0:30], out)
mape = svr.calc_mape(out, testtarget[0:30])
predx = svr.calc_pred(out, testtarget[0:30], 25)
rsq = svr.calc_rsqr(out, testtarget[0:30])
print out
print testtarget[0:30]
# print model results!
x = array(testinput[0:30], dtype=int32)
y = array(testtarget[0:30], dtype=int32)
xp = array(testinput[0:30], dtype=int32)
yp = array(out, dtype=int32)
fig = figure()
ax1 = fig.add_subplot(1,1,1)
ax1.title.set_text("Predizioni modello SVR con C= %f, Gamma = %f, Eps = %f" % (C, gamma, eps))
realvalues = ax1.plot(x, y)
predictedvalues = ax1.plot(xp,yp,"r")
ax1.axis([8.9,max(xp)+0.5,0,max(y)+10])
ax1.set_xlabel('minutes of the week')
ax1.set_ylabel('number of requests')
legend([realvalues,predictedvalues], ["Real Values","Predicted Values"])
fig.savefig("svr_model_%f" % time(), format='png')
print "SME = %f" % sme
print "MAPE = %f" % mape
print "R^2 = %f" % rsq
print "PREDX = %f" % predx
def hmm(states_nuber):
#initialization of data wmproxy
# traininput, traintarget, testinput, testtarget = initialize_wmproxy()
#initialization of EWS data
traininput, traintarget, pred_test, testinput, testtarget = initialize_ews()
## In this case we will try out performance of HMM considering just Monday! We will concatenate all series of data representing Monday workload!
## With EWS service we have three weeks as training and one week as test
# trainelements = []
# traintarget_new = zip(*traintarget[0])
#
# for mon in traintarget_new:
# trainelements += mon
#
# trainelements = log(trainelements)
#
# print "Monday training = %s" % trainelements
model = HMM(traintarget, states_nuber, 264)
test = pred_test[0:4]
# test = traintarget[1370:1409]
# for i in range(len(testtarget)):
# if(testtarget[i] != 0):
# testtarget[i] = numpy.log(testtarget[i])
# else:
# testtarget[i] = 1.0/100000000000
#
#
# This function predict a timewindow X startinf from a test sequence
states = model.hmm_req(test, 30)
# This function is for the integrity check of the model (how it fit the training set)
# seq = EmissionSequence(model.sigma, traintarget)
# states = model.m.viterbi(seq)
# states = states[0]
ttarget = []
print "States2"
print states
meanout = []
for state in states:
li = model.m.getEmission(state)
# maxes = numpy.where(array(li) > max(li)*0.4)[0]
# print maxes
# meanout.append(li.index(li[max(maxes)]))
# maxvals = [li.index(li[maxval]) for maxval in maxes]
maxes = nlargest(5, li)
meanout.append(li.index(maxes[0]))
maxvals = [li.index(maxval) for maxval in maxes]
ttarget.append(maxvals)
## sme = sme_calc(ttarget, testtarget[counter])
# print ttarget
#
minout = []
maxout = []
for element in ttarget:
minout.append(min(element))
maxout.append(max(element))
print len(meanout)
# print "minout %s: " % minout
# print "meanout %s: " % meanout
# print "maxout %s: " % maxout
x = array(traininput[0:30], dtype=int32)
y = array(pred_test[5:35], dtype=int32)
xp = array(traininput[0:30], dtype=int32)
yp = array(minout, dtype=int32)
xp1 = array(traininput[0:30], dtype=int32)
yp1 = array(maxout, dtype=int32)
xp2 = array(traininput[0:30], dtype=int32)
yp2 = array(meanout, dtype=int32)
fig = figure()
print "len x = % d" % len(x)
print "len y = % d" % len(y)
ax1 = fig.add_subplot(1,1,1)
ax1.title.set_text("Predizioni modello HMM con %d stati" % (states_nuber))
realvalues = ax1.plot(x, y)
minpred = ax1.plot(xp,yp,"r")
maxpred = ax1.plot(xp1,yp1,"g")
avgpred = ax1.plot(xp2,yp2,"y")
# ax1.axis([8.9,max(xp)+0.5,0,max(y)+10])
ax1.set_xlabel('minutes of the week')
ax1.set_ylabel('cluster')
legend([realvalues, minpred, avgpred, maxpred], ["Real Values", "Minimum Predicted Values","Average Predicted Values","Maximum Predicted Values"])
# fig.savefig("hmm_model_%f.png" % time(), format='png')
# sme = model.sme_calc(ttarget, testtarget[10:30])
# mape = model.mape_calc(ttarget, testtarget[10:30])
# predx = model.pred_calc(ttarget, testtarget[10:30], 25)
# rsq = model.rsqr_calc(ttarget, testtarget[10:30])
#
# print "SME = %f" % sme
# print "MAPE = %f" % mape
# print "R^2 = %f" % rsq
# print "PREDX = %f" % predx
# Compute the error of the on the worst case and the most probable value of the probabilities
max_error = []
mean_error = []
for i in range(len(maxout)):
max_error.append(maxout[i] - traintarget[i])
mean_error.append(meanout[i] - traintarget[i])
fig2 = figure()
ax = fig2.add_subplot(1,1,1)
ax.title.set_text("Errore rispetto alle predizioni: worst case e most probable (%d Stati)" % (states_nuber))
maxerr = ax.plot(x,array(max_error, dtype=int32))
meanerr = ax.plot(x,array(mean_error, dtype=int32))
legend([maxerr, meanerr], ["WC error", "MP error"])
print "Mean WC error"
print mean(max_error)
print "Mean WC error (abs)"
print mean(absolute(max_error))
print "Mean MP error"
print mean(mean_error)
print "Mean MP error (abs)"
print mean(absolute(mean_error))
print "Number of underestimations"
print len(numpy.where(array(max_error) < 0)[0])
return model, pred_test
def mcmc():
from thesis.scripts.bayesian import requestModel_nocl
model = MCMC(requestModel_nocl)
traintarget = model.traintarget
testtarget = model.testtarget
traininput = model.traininput
testinput = model.testinput
starttime = time()
iter = 1000
model.sample(iter=iter, burn=200, thin=10)
print "Training time"
print time() - starttime
for i in range(len(testinput)):
testinput[i] -= 10080
reqs = poisson_req_nocl(model, testinput[0:30], testtarget[0:30])
ttarget = []
for prob in reqs:
m = max(prob)
el = pylab.find(prob > m*2/3)
maxes = nlargest(15, prob)
maxvals = [prob.index(maxval) for maxval in maxes]
ttarget.append(maxvals)
# sme = sme_calc_nocl(ttarget, testtarget[0:20])
# mape = mape_calc(ttarget, testtarget[0:20])
# predx = pred_calc(ttarget, testtarget[0:20], 0.25)
# rsq = rsqr_calc(ttarget, testtarget[0:20])
#
# print "SME = %f" % sme
# print "MAPE = %f" % mape
# print "R^2 = %f" % rsq
# print "PREDX = %f" % predx
minout = []
maxout = []
meanout = []
for element in ttarget:
minout.append(min(element))
maxout.append(max(element))
meanout.append(mean(element))
print "minout %s: " % minout
print "meanout %s: " % meanout
print "maxout %s: " % maxout
x = array(testinput[0:30], dtype=int32)
y = array(testtarget[0:30], dtype=int32)
xp = array(testinput[0:30], dtype=int32)
yp = array(minout, dtype=int32)
xp1 = array(testinput[0:30], dtype=int32)
yp1 = array(maxout, dtype=int32)
xp2 = array(testinput[0:30], dtype=int32)
yp2 = array(meanout, dtype=int32)
fig = figure()
ax1 = fig.add_subplot(1,1,1)
ax1.title.set_text("Predizioni modello MCMC con %d iterazioni" % (iter))
realvalues = ax1.plot(x, y)
minpred = ax1.plot(xp,yp,"r")
maxpred = ax1.plot(xp1,yp1,"g")
avgpred = ax1.plot(xp2,yp2,"y")
# ax1.axis([8.9,max(xp)+0.5,0,max(y)+10])
ax1.set_xlabel('minutes of the week')
ax1.set_ylabel('number of requests')
legend([realvalues,minpred, avgpred, maxpred], ["Real Values","Minimum Predicted Values","Average Predicted Values","Maximum Predicted Values"])
fig.savefig("mcmc_model_%f" % time(), format='png')
subplot(111, xscale="log")
xlabel("Episodes")
ylabel("Performance")
legends = []
for k in xrange(nPerfFiles):
# print "- ",sys.argv[1+k*2]
legends.append(sys.argv[1 + k * 2])
lines = open(sys.argv[2 + k * 2], "r").readlines()
# Remove commented lines
while lines[0][0] == "#":
del lines[0]
# Then, the first line contains the total number of episodes
# after each iteration
strNEpis = lines.pop(0).split(" ")
nIters = len(strNEpis)
nEpis = empty(nIters)
for i in xrange(nIters):
nEpis[i] = float(strNEpis[i])
# Now we can go through the perf of each trial and each iteration
nTrials = len(lines)
perf = empty((nTrials, nIters))
for i in xrange(nTrials):
sp = lines[i].split(" ")
for j in xrange(nIters):
perf[i, j] = float(sp[j])
# plot(nEpis,mean(perf, axis=0))
errorbar(nEpis, mean(perf, axis=0), yerr=std(perf, axis=0), label=sys.argv[1 + k * 2])
# legend(legends,loc='best')
legend(loc="best")
show()
def GetVelFromQ(self,el):
'''
Computing velocity profile in terms of the flow rate,
using inverse womersley method of Cezeaux et al.1997
'''
self.radius = mean(el.Radius)
self.Res = el.R
self.length = el.Length
self.Name = el.Name
Flow = mean(self.signal)
#WOMERSLEY NUMBER
self.alpha = self.radius * sqrt((2.0 *pi*self.density)/(self.tPeriod*self.viscosity))
self.Wom = self.alpha
self.Re = (2.0*Flow*self.SimulationContext.Context['blood_density'])/(pi*self.radius*self.SimulationContext.Context['dynamic_viscosity'])
#FOURIER SIGNAL
k = len(self.signal)
n = 0
while n < (self.nHarmonics):
An = 0
Bn = 0
for i in arange(k):
An += self.signal[i] * cos(n*(2.0*pi/self.tPeriod)*self.dt*self.nSteps[i])
Bn += self.signal[i] * sin(n*(2.0*pi/self.tPeriod)*self.dt*self.nSteps[i])
An = An * (2.0/k)
Bn = Bn * (2.0/k)
self.fourierModes.append(complex(An, Bn))
n+=1
self.fourierModes[0] *= 0.5 #mean Flow, as expected. It's defined into xml input file.
self.Steps = linspace(0,self.tPeriod,self.samples)
self.VelRadius = {}
self.VelRadiusSteps = {}
self.VelocityPlot = {}
for step in self.Steps:
self.Velocity = {}
y = -1 # raggio da -1 a 1, 200 punti.
while y <=1.:
self.VelRadius[y] = 2*(1.0**2 - y**2)*self.fourierModes[0]
y+=0.01
k=1
while k < self.nHarmonics:
cI = complex(0.,1.)
cA = (self.alpha * pow((1.0*k),0.5)) * pow(cI,1.5)
c1 = 2.0 * jn(1, cA)
c0 = cA * jn(0, cA)
cT = complex(0, -2.0*pi*k*self.t/self.tPeriod)
y=-1 #da -1 a 1 #y=0 #centerline
while y<=1.0:
'''vel computation'''
c0_y = cA * jn(0, (cA*y))
vNum = c0-c0_y
vDen = c0-c1
vFract = vNum/vDen
cV = self.fourierModes[k] * exp(cT) * vFract
self.VelRadius[y] += cV.real #valore di velocity riferito al raggio adimensionalizzato
self.Velocity[y] = self.VelRadius[y].real
y+=0.01
k+=1
unsortedRadii = []
for rad, vel in self.Velocity.iteritems():
unsortedRadii.append(rad)
radii = sorted(unsortedRadii)
self.VelPlot = []
for x in radii:
for rad, vel in self.Velocity.iteritems():
if x == rad:
self.VelPlot.append(vel*(100.0/(self.radius**2*pi)))
self.VelocityPlot[step] = self.VelPlot
self.t += self.dtPlot
def getRootMeanSquaredErrors(calculated,real):
d = calculated-real
d = d**2
m = mean(d)
sm = sqrt(m)
return sm
statList = []
learningCurveStatList = []
for dataset in datasetNames:
print dataset
(data, labels) = readData(dataDir, dataset)
for algorithm in algorithmList:
print algorithm.__name__, '\t',
accuracyList = []
#the learning curve list has one list of values for each training set size
#so if we're trying a training set of size 10, learningCurveList[10] will be a
#list of the accuracies from the test with training size 10
learningCurveList = []
for x in range(len(labels)):
learningCurveList.append([])
crossValidation(numCrossValidationFolds, data, labels, algorithm, accuracyList, learningCurveList, numLearningCurveIterations, learningCurveIndexMod)
statList.append((dataset, algorithm.__name__, mean(accuracyList), std(accuracyList)))
learningCurveStatList.append((dataset, algorithm.__name__,
[mean(x) for x in learningCurveList],
[std(x) for x in learningCurveList]))
outFile = open(path.join(figureDir, "table.txt"), 'a')
outFile.write('algorithm')
for ds in datasetNames:
outFile.write('&& ' + ds)
outFile.write('\\\\ \n')
for alg in [x.__name__ for x in algorithmList]:
outFile.write(alg)
for ds in datasetNames:
relevantTupList = [x for x in statList if x[0] == ds and x[1] == alg]
if len(relevantTupList) != 1: