def segment_driver(driver_id):
''' this generated the segments in settings.SEGMENTS_FOLDER[1] '''
da = DataAccess()
for ride_id_minus_1, ride in enumerate(da.get_rides(driver_id)):
ride_id = ride_id_minus_1 + 1
if da.skip_segment(driver_id, ride_id):
continue
# apply the Ramer-Douglas-Peucker algorithm
ride = [p + [i] for i, p in enumerate(smoothen(ride))] # enrich with timestamp
ride = rdp(ride, epsilon=10)
lengths = [util.euclidian_distance(ride[i-1], ride[i]) for i in xrange(1, len(ride))]
times = [ride[i][2] - ride[i-1][2] for i in xrange(1, len(ride))]
angles = [util.get_angle(ride[i-2], ride[i-1], ride[i]) for i in xrange(2, len(ride))]
# bucket the values
lengths = util.bucket(np.log(lengths), 25, [2.2,8]) # [int(l) for l in lengths]
times = util.bucket(np.log(times), 20, [1,5.5]) # [int(t) for t in times]
angles = util.bucket(angles, 30, [0,180]) # [int(a) for a in angles]
# write results
da.write_ride_segments(driver_id, ride_id, lengths, times, angles)
logging.info('finished segmenting driver %s' % driver_id)
def segment_driver(driver_id):
''' this generated the segments in settings.SEGMENTS_FOLDER[1] '''
da = DataAccess()
for ride_id_minus_1, ride in enumerate(da.get_rides(driver_id)):
ride_id = ride_id_minus_1 + 1
if da.skip_segment(driver_id, ride_id):
continue
# apply the Ramer-Douglas-Peucker algorithm
ride = [p + [i]
for i, p in enumerate(smoothen(ride))] # enrich with timestamp
ride = rdp(ride, epsilon=10)
lengths = [
util.euclidian_distance(ride[i - 1], ride[i])
for i in xrange(1, len(ride))
]
times = [ride[i][2] - ride[i - 1][2] for i in xrange(1, len(ride))]
angles = [
util.get_angle(ride[i - 2], ride[i - 1], ride[i])
for i in xrange(2, len(ride))
]
# bucket the values
lengths = util.bucket(np.log(lengths), 25,
[2.2, 8]) # [int(l) for l in lengths]
times = util.bucket(np.log(times), 20,
[1, 5.5]) # [int(t) for t in times]
angles = util.bucket(angles, 30, [0, 180]) # [int(a) for a in angles]
# write results
da.write_ride_segments(driver_id, ride_id, lengths, times, angles)
logging.info('finished segmenting driver %s' % driver_id)
def run_model():
data, true_labels = ldl.get_data_linear()
true_buckets = [util.bucket(t) for t in true_labels]
data = np.tile(data, (DATA_MULTIPLIER, 1))
print("DATA SHAPE:", data.shape)
true_buckets = np.tile(true_buckets, DATA_MULTIPLIER)
# tuples of (batch_id, total regret, error while training, eval error, precision, recall)
batch_results = []
for T in range(NUM_BATCHES):
model = Lin_UCB(ALPHA)
#model = LASSO_BANDIT()
if False:
data, true_labels, columns_dict, values_dict = dl.get_data()
true_buckets = [util.bucket(t) for t in true_labels]
#model = Fixed_Dose(columns_dict, values_dict)
#model = Warfarin_Clinical_Dose(columns_dict, values_dict)
#model = Warfarin_Pharmacogenetic_Dose(columns_dict, values_dict)
batch_id = str(random.randint(100000, 999999))
print()
print("Start Batch: ", batch_id)
zipped_data = list(zip(data, true_buckets))
random.shuffle(zipped_data)
data, true_buckets = zip(*zipped_data)
data = np.array(data)
model.train(data, true_buckets)
pred_buckets = model.evaluate(data)
print(batch_id, "Performance on " + str(model))
acc, precision, recall = util.evaluate_performance(
pred_buckets, true_buckets)
print("\tAccuracy:", acc)
print("\tPrecision:", precision)
print("\tRecall:", recall)
plot_regret(model.regret, ALPHA, batch_id)
plot_error_rate(model.error_rate, ALPHA, batch_id)
batch_results.append(
(batch_id, model.get_regret()[-1], model.get_error_rate()[-1],
1 - acc, precision, recall))
with open('batch/regret' + str(model) + batch_id, 'wb') as fp:
pickle.dump(model.regret, fp)
with open('batch/error' + str(model) + batch_id, 'wb') as fp:
pickle.dump(model.error_rate, fp)
return batch_results
def evaluate_datum(self, datum):
dose = 5.6044
dose -= 0.2614 * datum[self.columns_dict['Age']]
dose += 0.0087 * datum[self.columns_dict['Height (cm)']]
dose += 0.0128 * datum[self.columns_dict['Weight (kg)']]
vk_gene = 'VKORC1 genotype: -1639 G>A (3673); chr16:31015190; rs9923231; C/T'
vk_gene2 = 'VKORC1 QC genotype: -1639 G>A (3673); chr16:31015190; rs9923231; C/T'
dose -= 0.8677 * (datum[self.columns_dict[vk_gene]]
== self.values_dict[vk_gene]['A/G']
or datum[self.columns_dict[vk_gene2]]
== self.values_dict[vk_gene2]['A/G'])
dose -= 1.6974 * (datum[self.columns_dict[vk_gene]]
== self.values_dict[vk_gene]['A/A']
or datum[self.columns_dict[vk_gene2]]
== self.values_dict[vk_gene2]['A/A'])
dose -= 0.4854 * (datum[self.columns_dict[vk_gene]]
== self.values_dict[vk_gene]['NA']
and datum[self.columns_dict[vk_gene2]]
== self.values_dict[vk_gene2]['NA'])
dose -= 0.5211 * datum[
self.columns_dict['CYP2C9 consensus']] == self.values_dict[
'CYP2C9 consensus']['*1/*2']
dose -= 0.9357 * datum[
self.columns_dict['CYP2C9 consensus']] == self.values_dict[
'CYP2C9 consensus']['*1/*3']
dose -= 1.0616 * datum[
self.columns_dict['CYP2C9 consensus']] == self.values_dict[
'CYP2C9 consensus']['*2/*2']
dose -= 1.9206 * datum[
self.columns_dict['CYP2C9 consensus']] == self.values_dict[
'CYP2C9 consensus']['*2/*3']
dose -= 2.3312 * datum[
self.columns_dict['CYP2C9 consensus']] == self.values_dict[
'CYP2C9 consensus']['*3/*3']
dose -= 0.2188 * datum[self.columns_dict[
'CYP2C9 consensus']] == self.values_dict['CYP2C9 consensus']['NA']
dose -= 0.1092 * (datum[self.columns_dict['Race']]
== self.values_dict['Race']['Asian'])
dose -= 0.2760 * (
datum[self.columns_dict['Race']]
== self.values_dict['Race']['Black or African American'])
dose -= 0.1032 * (datum[self.columns_dict['Race']]
== self.values_dict['Race']['NA'])
dose += 1.1816 * self._get_enzyme_inducer_status(datum)
#Enzyme inducer status
dose -= 0.5503 * datum[self.columns_dict['Amiodarone (Cordarone)']]
# dose calculated in appx.pdf states that it's the sqrt of weekly
return util.bucket(dose**2)
def evaluate_datum(self, datum):
dose = 4.0376
dose -= 0.2546 * datum[self.columns_dict['Age']]
dose += 0.0118 * datum[self.columns_dict['Height (cm)']]
dose += 0.0134 * datum[self.columns_dict['Weight (kg)']]
dose -= 0.6752 * (datum[self.columns_dict['Race']]
== self.values_dict['Race']['Asian'])
dose += 0.4060 * (
datum[self.columns_dict['Race']]
== self.values_dict['Race']['Black or African American'])
dose += 0.0443 * (datum[self.columns_dict['Race']]
== self.values_dict['Race']['NA'])
dose += 0.0443 * (datum[self.columns_dict['Race']]
== self.values_dict['Race']['Unknown'])
dose += 1.2799 * self._get_enzyme_inducer_status(datum)
dose -= 0.5695 * (datum[self.columns_dict['Amiodarone (Cordarone)']]
== self.values_dict['Amiodarone (Cordarone)']['1'])
# dose calculated in appx.pdf states that it's the sqrt of weekly
return util.bucket(dose**2)
for i in range(len(data)):
labels[i] = self._evaluate_datum(data[i])
return labels
def _evaluate_datum(self, features):
action_t = -1
best_reward = float('-inf')
for arm in range(self.K):
predicted_arm_reward = self._predict_reward(arm, features)
if predicted_arm_reward > best_reward:
action_t = arm
best_reward = predicted_arm_reward
assert action_t != -1, "[eval datum] No arm was selected..."
return action_t
# probably do not run this for decent results (below data isn't randomized)
# execute 'run_batches.py' instead on the lasso bandit model.
if __name__ == '__main__':
data, true_labels = ldl.get_data_linear()
true_buckets = [util.bucket(t) for t in true_labels]
lasso_bandit = LASSO_BANDIT()
lasso_bandit.train(data, true_buckets)
pred_buckets = lasso_bandit.evaluate(data)
acc = util.get_accuracy_bucketed(pred_buckets, true_buckets)
print("accuracy on LASSO bandit: " + str(acc))
#plot_regret(lasso_bandit.regret, ALPHA)
#plot_error_rate(lasso_bandit.error_rate, ALPHA)
def evaluate_datum(self, datum):
return util.bucket(35)
