def dataReceived(self,data):
data = data.strip("\r\n")
if data.startswith("vote"):
self.factory.datastore.vote(data.split("|")[1])
elif data.startswith("add"):
args = data.split("|")
self.factory.datastore.add(args[1],args[2]) #name, path
def run(self, clf):
X, y = data.split(self.data)
skf_config = config.xvalidation
self.skf = StratifiedShuffleSplit(y, **skf_config)
self.X = X
self.y = y
report = Report(" | ".join([step[0] for step in clf.steps]))
X, y = self.X, self.y
y_true = []
y_predicted = []
for train_index, test_index in self.skf:
X_train, y_train = X[train_index], y[train_index]
X_test, y_test = X[test_index], y[test_index]
y_true += list(y_test)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
y_predicted += list(y_pred)
report.summary = classification_report(y_true,
y_predicted,
target_names=("dead", "alive"))
cm = confusion_matrix(y_true, y_predicted)
report.confusion_matrix = cm
report.confusion_matrix_norm = cm.astype("float") / cm.sum(
axis=1)[:, np.newaxis]
report.accuracy = accuracy_score(y_true, y_predicted)
return report
def run(self, clf):
X, y = data.split(self.data)
skf_config = config.xvalidation
self.skf = StratifiedShuffleSplit(y, **skf_config)
self.X = X
self.y = y
report = Report(" | ".join([step[0] for step in clf.steps]))
X, y = self.X, self.y
y_true = []
y_predicted = []
for train_index, test_index in self.skf:
X_train, y_train = X[train_index], y[train_index]
X_test, y_test = X[test_index], y[test_index]
y_true += list(y_test)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
y_predicted += list(y_pred)
report.summary = classification_report(y_true, y_predicted, target_names=("dead", "alive"))
cm = confusion_matrix(y_true, y_predicted)
report.confusion_matrix = cm
report.confusion_matrix_norm = cm.astype("float") / cm.sum(axis=1)[:, np.newaxis]
report.accuracy = accuracy_score(y_true, y_predicted)
return report
def test_split():
indices = np.array([2, 4, 8, 9])
expected = np.concatenate((indices, indices + 10))
actual = np.array(list(data.split(indices, 10, range(20))))
assert len(actual) == len(expected)
assert actual.dtype == expected.dtype
assert all(actual == expected)
def main(file_size):
"""Generates the files needed to run tagWeighting.py (the locality script).
"""
data_funcs_by_size = {
'small': data.get_small,
'medium': data.get_medium,
'large': data.get_large
}
all_data = data_funcs_by_size[file_size]()
train, test = data.split(all_data, 0.8)
metadata = []
photo = []
uid = []
for img in train:
pictureID = img['watchlink'].strip()
tags = img['tags']
lat, lon = img['latitude'], img['longitude']
userID = img['userID']
metadata.append('{0}\t0\t0\t{1}'.format(pictureID, ', '.join(tags)))
photo.append('0\t0\t{0}\t{1}'.format(lon, lat))
uid.append('0\t{0}'.format(userID))
write_to_file(metadata, file_size + '_train_metadata')
write_to_file(photo, file_size + '_train_photo')
write_to_file(uid, file_size + '_train_uid')
def process(self, opType):
if(len(self.threads) == 0):
self.responseLine.setText("No connection.")
return
#Get data inputed on UI
dt = self.vectorLine.text()
if(len(dt) == 0):
self.responseLine.setText("No data to proccess.")
return
#Creates sub-vectors
dt = list(map(int, dt.split()))
vectorOfSubvectors = self.vectorFragmentation(dt)
resp = dt[0]
if opType == "MAX":
typeFunction = max
elif opType == "MIN":
typeFunction = min
else:
self.responseLine.setText("Type of function error")
self.update()
return
for i in range(len(self.threads)):
resp = typeFunction(resp, int(self.threads[i].sendData(opType + "#" + " ".join(str(x) for x in vectorOfSubvectors[i]))))
self.responseLine.setText(opType + " VALUE: " + str(resp))
self.update()
def save_layout_file(name, view, layout, controls):
dir = java.io.File('layouts')
make_layout_dir(dir)
fn = 'layouts/%s.layout' % name
# Check if file exists
# - If it does, extract java layout information, otherwise just make a new file
java_layout = ""
if java.io.File(fn).exists():
# Save a backup of the layout file
fp = java.io.File(fn).getCanonicalPath()
copyfile(fp, fp + '.bak')
f = file(fn, 'r')
data = f.read()
for line in data.split('\n'):
if line.startswith('#'):
java_layout += '\n' + line
f.close()
f = file(fn, 'w')
layout_text = ',\n '.join([repr(x) for x in layout])
f.write('(%s,\n [%s],\n %s) %s' % (view, layout_text, controls, java_layout))
f.close()
def load_saved_lf(filename='lf_data.dat'):
lf_lines = open(filename, 'r').readlines()
dic = {}
for line in lf_lines:
if line.startswith('#'): continue
line = line.strip()
filter = line.split(' ')[0]
region = line.split(' ')[1]
data = line[7:]
bins = map(float, data.split(';')[0].split())
hist = map(float, data.split(';')[1].split())
if not filter in dic.keys(): dic[filter] = {}
if not region in dic[filter].keys(): dic[filter][region] = {}
#if not 'bins' in dic[filter][region].keys(): dic[filter][region]['bins'] = {}
dic[filter][region] = {'bins': bins, 'hist': hist}
return dic
def save_layout_file(name, view, layout, controls):
dir = java.io.File('layouts')
make_layout_dir(dir)
fn = 'layouts/%s.layout' % name
# Check if file exists
# - If it does, extract java layout information, otherwise just make a new file
java_layout = ""
if java.io.File(fn).exists():
# Save a backup of the layout file
fp = java.io.File(fn).getCanonicalPath()
copyfile(fp, fp + '.bak')
f = file(fn, 'r')
data = f.read()
for line in data.split('\n'):
if line.startswith('#'):
java_layout += '\n' + line
f.close()
f = file(fn, 'w')
layout_text = ',\n '.join([repr(x) for x in layout])
f.write('(%s,\n [%s],\n %s) %s' %
(view, layout_text, controls, java_layout))
f.close()
def main():
print("load csv into a pandas dataframe")
dt = data_loader()
data = dt.load_data()
print(f"data has {data.shape}")
data = dt.encode_target(data)
print(
"preprocess data by removing outliers and encoding feature variables")
data = dt.preprocess(data)
#print(data.columns)
print(
"scale data using standardscaler and encoding using pandas get_dummies"
)
data = dt.scale_columns(data)
print(f"data contains {data.columns}")
sam = resample_data()
data = sam.under_sample(data)
print(data['y'].value_counts())
s = split()
data = s.train_test(data)
print(data[0].shape)
classifiers_cv(data[0], data[1], data[2], data[3])
def parse_to_dict(data):
programs = {}
for line in data.split('\n'):
key, values = line.split(' <-> ')
key = int(key)
values = tuple((int(v) for v in values.split(',')))
programs[key] = values
return programs
def parse_text(data):
parts = []
for pipe in data.split('\n'):
a, b = pipe.split('/')
parts.append(Pipe(int(a), int(b)))
return Parts(parts)
def train_test_split(self, X, y, eval_size):
if eval_size:
X_train, X_valid, y_train, y_valid = data.split(
X, y, test_size=eval_size)
else:
X_train, y_train = X, y
X_valid, y_valid = X[len(X):], y[len(y):]
return X_train, X_valid, y_train, y_valid
def train_test_split(self, X, y, eval_size):
if eval_size:
X_train, X_valid, y_train, y_valid = data.split(
X, y, test_size=eval_size)
else:
X_train, y_train = X, y
X_valid, y_valid = X[len(X):], y[len(y):]
return X_train, X_valid, y_train, y_valid
def compare_lenght(in_path):
seq_dict = imgs.read_seqs(in_path)
len_dict = get_len_dict(seq_dict)
train, test = data.split(len_dict.keys())
train, test = by_cat(train), by_cat(test)
for cat_i in train.keys():
train_i = np.mean([len_dict[name_i] for name_i in train[cat_i]])
test_i = np.mean([len_dict[name_i] for name_i in test[cat_i]])
print("%d,%.2f,%.2f" % (cat_i, test_i, train_i))
def test():
clf = RandomForest
X_train, y_train = data.split(data.train)
X_test = data.extract_features(data.test)
clf.fit(X_train, y_train)
labels = clf.predict(X_test)
pd.DataFrame({
"PassengerId": np.array(data.test["PassengerId"]),
"Survived": labels
}).to_csv("submit.csv", index=False)
def person_model(in_path, out_path, n_epochs=100):
seq_dict = imgs.read_seqs(in_path)
train, test = data.split(seq_dict.keys())
persons = [data.parse_name(name_i)[1] - 1 for name_i in train]
persons = keras.utils.to_categorical(persons)
X, y = to_dataset(train, seq_dict)
n_cats, n_channels = y.shape[1], X.shape[-1]
model = models.make_exp(n_cats, n_channels)
model.summary()
model.fit(X, y, epochs=n_epochs, batch_size=256)
model.save(out_path)
def test():
clf = RandomForest
X_train, y_train = data.split(data.train)
X_test = data.extract_features(data.test)
clf.fit(X_train, y_train)
labels = clf.predict(X_test)
pd.DataFrame({
"PassengerId": np.array(data.test["PassengerId"]),
"Survived": labels
}).to_csv("submit.csv", index=False)
def agum_template(raw_path,agum_path,agum,n_iters=10):
raw_data=imgs.read_seqs(raw_path)
train,test=data.split(raw_data.keys())
train_data={ name_i:raw_data[name_i] for name_i in train}
agum_dict={}
for name_i,seq_i in list(train_data.items()):
agum_seq_i = agum(images=seq_i)
for j in range(n_iters):
new_name_i=name_i+'_'+str(j)
print(new_name_i)
agum_dict[new_name_i]=agum_seq_i
new_dict={**raw_data,**agum_dict}
imgs.save_seqs(new_dict,agum_path)
def apply_physics_properties_to_shape( self, node, shape ):
shape.restitution = DEFAULT_RESTITUTION
shape.density = DEFAULT_DENSITY
shape.friction = DEFAULT_FRICTION
data = node.getAttribute('physics_shape' )
if data:
keyvalues = data.split(',')
for keyvalue in keyvalues:
key,value = keyvalue.split('=')
value = self.cast_value( value )
setattr( shape, key, value )
def load_data(in_path, split=True):
feat_dict = single.read_frame_feats(in_path)
if (split):
train, test = data.split(feat_dict.keys())
train, test = prepare_data(train,
feat_dict), prepare_data(test, feat_dict)
params = {
'ts_len': train[0].shape[1],
'n_feats': train[0].shape[2],
'n_cats': train[1].shape[1]
}
return train, test, params
else:
names = list(feat_dict.keys())
return prepare_data(names, feat_dict), names
def run(self, msg, user):
c = httplib2.HTTPSConnectionWithTimeout("api.mrbesen.de")
c.request("GET", "/pos.php")
response = c.getresponse()
if response.code == 200:
data = response.read().decode().strip()
time, loc, lastknowntime, lastknownloc = data.split("\n")
if loc == 'unknown':
print("unknown location")
datestr = self.getTimeStr(time)
return "Also am " + datestr + " war ich in " + lastknownloc + " aber grade bin ich Unterwegs."
else:
datestr = self.getTimeStr(time)
return "Bin gerade (" + datestr + ") in " + loc
return None
def sim_model(in_path,out_path):
full=img_dataset(in_path)
train,test=data.split(full.keys())
# train={ name_i:full[name_i] for name_i in train}
X0,X1,y=[],[],[]
for i,name_i in enumerate(train):
for name_j in train[i:]:
y_k=int(name_i.split("_")[0]==name_j.split("_")[0])
X0.append( full[name_i])
X1.append(full[name_j])
y.append(y_k)
X=[np.array(X0),np.array(X1)]
sim_metric,model=sim.frames.make_five(20,1)
sim_metric.fit(X,y,epochs=250,batch_size=100)
if(out_path):
model.save(out_path)
def get_data():
global times
global tidalVolume
global oldtidalVolume
global peakPressure
global oldpeakPressure
global respirationRate
global oldrespirationRate
global oldTime
global maxTime
global corrupt
global PEEP
maxTime = 0
startingTime = time.time()
while (corrupt and (int(time.time() - startingTime) <= 1)):
try:
ser = serial.Serial('COM3', baudrate)
corrupt = False
while True:
while (ser.inWaiting() == 0):
pass
value = ser.readline()
try:
data = str(value.decode("utf-8"))
data = data.split(",")
dataTime = int(data[0])
signal1 = int(data[1]) - 500
update_level(dataTime, 0, signal1, 0)
except:
pass
except:
pass
if corrupt == True:
currTime = 0.0
while True:
if currTime / 1000 not in data1.keys():
currTime = 0.0
maxTime = 0
#pp = data1[currTime/1000]/30
pp = 45
rr = data2[currTime / 1000] / 12
tv = data3[currTime / 1000] / 3
update_level(currTime, pp, rr, tv)
currTime += 10
time.sleep(0.01)
def get_data(file_size):
"""Fetches training and test data.
Args:
file_size: 'small', 'medium', or 'large' indicating the size of the desired dataset
Returns:
(train_data, test_data) where train_data and test_data are lists of data points (each data point is a dict)
"""
data_funcs_by_size = {
'small': data.get_small,
'medium': data.get_medium,
'large': data.get_large
}
all_data = data_funcs_by_size[file_size]()
train_data, test_data = data.split(all_data, 0.8)
return train_data, test_data
def _ParseCookie(self, cookie):
"""Parses the cookie and returns NULL_COOKIE if it's invalid.
Args:
cookie: The text of the cookie.
Returns:
A map containing the values in the cookie.
"""
try:
(hashed, data) = cookie.split('|', 1)
# global cookie_secret
if hashed != str(hash(cookie_secret + data) & 0x7FFFFFF):
return self.NULL_COOKIE
values = data.split('|')
return {
COOKIE_UID: values[0],
COOKIE_ADMIN: values[1] == 'admin',
COOKIE_AUTHOR: values[2] == 'author',
}
except (IndexError, ValueError):
return self.NULL_COOKIE
def _ParseCookie(self, cookie):
"""Parses the cookie and returns NULL_COOKIE if it's invalid.
Args:
cookie: The text of the cookie.
Returns:
A map containing the values in the cookie.
"""
try:
(hashed, data) = cookie.split('|', 1)
# global cookie_secret
if hashed != str(hash(cookie_secret + data) & 0x7FFFFFF):
return self.NULL_COOKIE
values = data.split('|')
return {
COOKIE_UID: values[0],
COOKIE_ADMIN: values[1] == 'admin',
COOKIE_AUTHOR: values[2] == 'author',
}
except (IndexError, ValueError):
return self.NULL_COOKIE
def load_params(train_prog, train_exe, place, logger, args=None):
if not args.para_load_dir:
return
logger.info('loading para from {}'.format(args.para_load_dir))
param_list = train_prog.block(0).all_parameters()
param_name_list = [p.name for p in param_list]
for data in listDir(args.para_load_dir):
slot = int(data.split('_')[1].split('.')[0])
with open(data, 'rb') as fin:
if six.PY2:
p_array = pickle.load(fin)
else:
p_array = pickle.load(fin, encoding='bytes')
p_array = p_array.reshape((-1))
offset = 0
for name in name_dict:
s = name_dict[name]
if s == slot:
card = 0
#for scope in [train_exe.scope]:#train_exe.executor.local_scopes():
for scope in train_exe.executor.local_scopes():
tensor = scope.find_var(name).get_tensor()
shape = tensor.shape()
tensor_len = np.prod(shape)
new_array = p_array[offset:offset + tensor_len]
new_array = new_array.reshape(shape)
if args.use_gpu:
placex = fluid.CUDAPlace(card)
else:
placex = fluid.CPUPlace()
tensor.set(new_array.astype(np.float32), placex)
logger.info('card {} loaded {}[{}] from {}[{}:{}]'.
format(card, name, shape, data, offset,
offset + tensor_len))
card = card + 1
offset += tensor_len
def cmd_tio(self, msg, args, stdin):
'''
Run {prefix}tio [language] [code] to evaluate code in a given language on
Try it online! (https://tio.run/). Specify additional sections on a
separate line consisting of three hashes (###) followed by the section
name, which can be any of: stdin (provide input), arg (provide any number
of command line arguments), or stderr (specify this section to view stderr
output in addition to stdout; you may also retroactively do this with
{prefix}tio err).
'''
err = " (try `{}help tio` for more information)".format(self.prefix)
if not args: return 'Basic usage: {}tio [lang] [code]'.format(self.prefix) + err
if args == 'err': return self.tioerr
lang, *rest = args.split(None, 1)
rest = rest[0] if len(rest) else ''
stdin = ''
stderr = False
args = []
code, *parts = rest.split('\n###')
for part in parts:
name, data = part.split(None, 1) if '\n' in part or ' ' in part else (part, '')
name = name.strip()
if name == 'stdin': stdin = data
elif name == 'stderr': stderr = True
elif name == 'arg': args.append(data)
else: return "Unknown section `{}`".format(name) + err
try:
data = requests.post('https://tio.run/cgi-bin/run/api/', zlib.compress(bytes('Vlang\u00001\u0000{}\u0000F.code.tio\u0000{}\u0000{}F.input.tio\u0000{}\u0000{}Vargs\u0000{}{}\u0000R'.format(lang, len(bytes(code, 'utf-8')), code, len(bytes(stdin, 'utf-8')), stdin, len(args), (len(args) * '\u0000{}').format(*args)), 'utf-8'), 9)[2:-4], timeout=5).content.decode('utf-8')
data = data.split(data[:16])[1:]
if len(data) == 1: return data[0] # error
dout, derr = [x.strip('\n') for x in data[:2]]
self.tioerr = derr
haserr = re.search('\nReal time: \\d+\\.\\d+ s\nUser time: \\d+\\.\\d+ s\nSys\\. time: \\d+\\.\\d+ s\nCPU share: \\d+\\.\\d+ %\nExit code: \\d+$', data[1]).start() > 0
return (dout+'\n--- stderr ---\n'+derr if stderr else dout+('\n[stderr output - use {}tio err to view]'.format(self.prefix) if haserr else '')) or '[no output]'
except requests.exceptions.ConnectionError:
return '5 second timeout reached.'
# digits_session_2_dataset
# digits_session_3_dataset
# digits_session_4_dataset
#
# sequence_groups[i] contains data for class i
# 4-dimensional data structure: (class, sequence_num, timestep, channel_num)
sequence_groups = transform_data(data.digits_session_4_dataset())
# Split sequence_groups into training and validation data
#training_sequence_groups, validation_sequence_groups = data.split(sequence_groups, 1./6)
# Manually selecting different training and validation datasets
training_sequence_groups = transform_data(data.digits_session_1_dataset())
validation_sequence_groups = transform_data(data.digits_session_4_dataset())
validation_sequence_groups = data.split(validation_sequence_groups, 1 / 6.)[1]
# Pads or truncates each sequence to length
length = 600
training_sequence_groups = data.transform.pad_truncate(
training_sequence_groups, length)
validation_sequence_groups = data.transform.pad_truncate(
validation_sequence_groups, length)
# Format into sequences and labels
train_sequences, train_labels = data.get_inputs(training_sequence_groups)
val_sequences, val_labels = data.get_inputs(validation_sequence_groups)
# Calculate sample weights
class_weights = compute_class_weight('balanced', np.unique(train_labels),
train_labels)
train_weights = class_weights[list(train_labels)]
#sequence_groups = data.combine([
# data.process(10, ['data/data/2_subvocal_digits_9_trials.txt']),
# data.process(10, ['data/data/3_subvocal_digits_11_trials.txt']),
# data.process(10, ['data/data/4_subvocal_digits_10_trials.txt']),
# data.process(10, ['data/data/5_subvocal_digits_10_trials.txt']),
# data.process(10, ['data/data/6_subvocal_digits_10_trials.txt']),
# ])
#sequence_groups = sequence_groups[:2]
sequence_groups = data.transform.default_transform(sequence_groups)
#sequence_groups = np.array(map(lambda x: x[30:], sequence_groups))
# Split into training and validation data
training_sequence_groups, validation_sequence_groups = data.split(sequence_groups, 1./6)
# Augment training data by varying positioning of padding/truncating
# Uncomment
#training_sequence_groups = data.transform.pad_extra(training_sequence_groups, length)
#training_sequence_groups = data.transform.augment_pad_truncate_intervals(training_sequence_groups, length, 10)
#training_sequence_groups = data.transform.augment_pad_truncate_intervals(training_sequence_groups, length, 100)
#training_sequence_groups = data.transform.augment_pad_truncate_intervals(training_sequence_groups, length, 50)
training_sequence_groups = data.transform.pad_truncate(training_sequence_groups, length)
#training_sequence_groups = data.transform.augment(training_sequence_groups,
# [(data.transform.gaussian_filter, [3], {})],
# include_original=True)
clf.fit(x_train,y_train)
# result = pd.DataFrame.from_dict(clf.cv_results_)
# with open(m[0]+'.csv','w') as f:
# result.to_csv(f)
print('The parameters of the best '+m[0]+' are: ')
print(clf.best_params_)
y_pred = clf.predict(x_train)
print(classification_report(y_true=y_train, y_pred=y_pred))
y_test_pred = clf.predict(x_test)
# print(classification_report(y_true=y_test, y_pred=y_test_pred))
# df_test_y = pd.DataFrame(y_test_pred , columns=['Survived'])
df = pd.DataFrame(data.get_test_PassengerId()).join(pd.DataFrame(y_test_pred , columns=['Survived']))
print(df.head())
df.to_csv('./titanic_test_result_'+m[0]+'.csv',index=False)
import data,preprocess
if '__main__' == __name__:
train_data = data.get_train_data()
train_data =preprocess.fill_missing_data(train_data ,sex_cat=True, embarked_one_hot=True)
# train_data = preprocess.feature_selection(train_data)
# train_data = preprocess.detect_outlier(train_data,drop=True)
print(train_data.head())
x_train,y_train = data.split(train_data)
# print(y_train.values)
x_test,y_test = data.get_test_x(),data.get_test_y()
x_test =preprocess.fill_missing_data(x_test,is_train=False,sex_cat=True, embarked_one_hot=True)
# poly = PolynomialFeatures(2,interaction_only=True)
# x_train = poly.fit_transform(x_train.values)
# x_test = poly.fit_transform(x_test.values)
model_selection(x_train.values,y_train.values,x_test.values,y_test.values)
def split_step(preprocessed_data_path, train_pct):
from data import split
train_data_path, test_data_path = split(preprocessed_data_path, train_pct)
return train_data_path, test_data_path
# Update internal cache of parameters
params['num_samples'] = num_samples
params['num_features'] = num_features
params['num_frames'] = num_frames
params['x_scaler'] = MinMaxScaler(feature_range=(-1, 1))
params['y_scaler'] = MinMaxScaler(feature_range=(-1, 1))
params['num_testcases'] = 2
X = empty((num_frames, num_timesteps, num_features))
Y = empty((num_frames, num_predictions))
for i in range(num_samples - num_timesteps):
X[i] = non_stationary[i:i + num_timesteps, ]
Y[i] = non_stationary[i + num_timesteps:i + num_timesteps +
num_predictions, 0]
Y
X_scaled = np.array(
[params['x_scaler'].fit_transform(X[i]) for i in range(X.shape[0])])
Y_scaled = params['y_scaler'].fit_transform(Y)
Y_scaled
X_train, Y_train, X_test, Y_test = data.split(X_scaled, Y_scaled,
params['num_testcases'])
Y_test
y_unscaled = params['y_scaler'].inverse_transform(Y_test)
y_undiff = data.inverse_diff(y_unscaled, a[-(params['num_testcases'] + 1):,
0:1])[-params['num_testcases']:]
y_undiff
print(data.recover_Ytest(Y_test, a, params))
# hi
N_CLASSES = 10
HIDDEN_SIZE = 256
CONV_L2 = 5e-4
FC_L2 = 1e-3
DROPOUT = 0.5
LR = 0.1
BATCH_SIZE = 128
EPOCHS = 300
TEST_TIME_K = 5
x_train, y_train, x_test, y_test = data.load_numpy()
x_train, y_train, _, _ = data.split(x_train, y_train, PERCENTAGE)
N_TRAIN_BATCHES = int(x_train.shape[0] / BATCH_SIZE)
N_TEST_BATCHES = int(x_test.shape[0] / BATCH_SIZE)
model = learning.cnn_classifier(number_of_classes=N_CLASSES,
hidden_layer_size=HIDDEN_SIZE,
conv_l2=CONV_L2,
fc_l2=FC_L2,
drop_out=DROPOUT)
opt = SGD(learning_rate=LR, momentum=0.9)
model.compile(loss='categorical_crossentropy',
optimizer=opt,
metrics=['accuracy'],
# data.combine([
# map(lambda x: x[:30], data.digits_session_dependence_1_dataset(channels=range(1, 8))),
# map(lambda x: x[:30], data.digits_session_dependence_2_dataset(channels=range(1, 8))),
# map(lambda x: x[:40], data.digits_session_dependence_3_dataset(channels=range(1, 8))),
# ]))
#
#print map(len, sequence_groups)
#lens = map(len, data.get_inputs(sequence_groups)[0])
#print min(lens), np.mean(lens), max(lens)
# Split sequence_groups into training and validation data
#training_sequence_groups, validation_sequence_groups = data.split(sequence_groups, 1./6)
# Manually selecting different training and validation datasets
training_sequence_groups, validation_sequence_groups = data.split(
data.digits_session_dependence_3_dataset(channels=range(1, 8)), 1. / 6)
training_sequence_groups = transform_data(
data.combine([
# map(lambda x: x[:30], data.digits_session_dependence_1_dataset(channels=range(1, 8))),
# map(lambda x: x[:30], data.digits_session_dependence_2_dataset(channels=range(1, 8))),
# map(lambda x: x[:40], data.digits_session_dependence_3_dataset(channels=range(1, 8))),
training_sequence_groups,
]))
validation_sequence_groups = transform_data(
data.combine([
# map(lambda x: x[:30], data.digits_session_dependence_1_dataset(channels=range(1, 8))),
# map(lambda x: x[:30], data.digits_session_dependence_2_dataset(channels=range(1, 8))),
# map(lambda x: x[:40], data.digits_session_dependence_3_dataset(channels=range(1, 8))),
validation_sequence_groups,
]))
import pandas as pd
from data import get_dataset, split
from detl.mydb import db_client
from svm import SVMClassifier, confusion_matrix, accuracy
with db_client().as_default():
digits = get_dataset()
X_train, X_test, y_train, y_test = split(digits)
classifier = SVMClassifier()
classifier.fit(X_train, y_train)
pred = classifier.predict(X_test)
print(confusion_matrix(y_test, pred).data)
print('Accuracy', accuracy(y_test, pred).data)
