def main():
dataset = data.load_eviction(dataset_type=DATASET_TYPE)
losses = []
accs = []
precs = []
recs = []
with open(FILENAME, 'w+') as f:
for i in range(NUM_TESTS):
print('*' * 80)
print('Round %s' % i)
dataset.split()
train(dataset)
predictions = predict(dataset)
loss, acc, prec, rec = evaluate(dataset.y_test, predictions)
losses.append(loss)
accs.append(acc)
precs.append(prec)
recs.append(rec)
loss_avg = np.array(losses).mean()
accs_avg = np.array(accs).mean()
precs_avg = np.array(precs).mean()
recs_avg = np.array(recs).mean()
print('Loss average: %s' % loss_avg)
print('Accuracy average: %s' % accs_avg)
print('Precision average: %s' % precs_avg)
print('Recall average: %s' % recs_avg)
output = '%s, %s, %s, %s, %s\n' % ('nn', loss_avg, accs_avg, precs_avg,
recs_avg)
f.write(output)
"""Preprocessing script to find most stable features, i.e. those that are
selective across many randomized trials.
"""
import pandas as pd
from sklearn.linear_model import RandomizedLogisticRegression
import data
if __name__ == '__main__':
dataset = data.load_eviction()
descriptions = pd.read_pickle('data/private/feature_codes_to_names.pck')
print('Data loaded.')
rlogistic = RandomizedLogisticRegression(normalize=True)
rlogistic.fit(dataset.X_train, dataset.y_train)
print('Model fitted.')
features = sorted(
zip(map(lambda x: round(x, 4), rlogistic.scores_),
dataset.X_train.columns))
print('Number of features:\t\t%s' % len(features))
nonzero_features = [(score, code) for score, code in features if score > 0]
print('Number of nonzero features:\t%s' % len(nonzero_features))
columns = []
for score, code in nonzero_features:
columns.append(code)
print('-' * 80)
"""Predict eviction response variable.
"""
import data
from predict import utils
from sklearn.svm import OneClassSVM
if __name__ == '__main__':
dataset_pos = data.load_pos_eviction()
dataset_neg = data.load_neg_eviction()
dataset_all = data.load_eviction()
# nu: The proportion of outliers we expect in our data.
model_pos = OneClassSVM(kernel='linear', nu=0.9)
model_pos.fit(dataset_pos.X_train)
model_neg = OneClassSVM(kernel='linear', nu=0.1)
model_neg.fit(dataset_neg.X_train)
predictions_pos = model_pos.predict(dataset_all.X_train)
predictions_neg = model_neg.predict(dataset_all.X_train)
# +1 is inlier, -1 is outlier. We want those who are evicted, to be +1
# and those who are not evicted to be 0.
# Outliers, those evicted, to be 1.
predictions_neg = (predictions_neg == -1).astype(int)
# Inliers, those evicted, to be 1.
predictions_pos = (predictions_pos == 1).astype(int)
def handcrafted():
dataset = data.load_eviction(dataset_type='handcrafted')
fname = '%s/results_HANDCRAFTED_dataset_OVER.csv' % DIR
test_runner.predict(MODELS, dataset, fname, N_TESTS, OVERSAMPLE)
def rlogistic():
dataset = data.load_eviction(dataset_type='rlogistic')
fname = '%s/results_RLOGISTIC_dataset_OVER.csv' % DIR
test_runner.predict(MODELS, dataset, fname, N_TESTS, OVERSAMPLE)
def pca():
dataset = data.load_eviction(dataset_type='pca')
fname = '%s/results_PCA_dataset_OVER.csv' % DIR
test_runner.predict(MODELS, dataset, fname, N_TESTS, OVERSAMPLE)
def full():
dataset = data.load_eviction()
fname = '%s/results_FULL_dataset_OVER.csv' % DIR
test_runner.predict(MODELS, dataset, fname, N_TESTS, OVERSAMPLE)
