def _get_x_y_report(data, jump=0.01,max_pos=None):
if max_pos==None:
max_pos=1
x = []
y = []
# build x and y for graph
idx = 0
x_pos = 0
while data[idx][0]<x_pos-jump/2:
idx+=1
while x_pos < max_pos:
y_data = []
# get all the data between [current_jump-jump/2, current_jump+jump/2[
while idx<len(data) and data[idx][0] >= x_pos-jump/2 and data[idx][0] < x_pos+jump/2:
y_data.append(data[idx])
idx+=1
if y_data:
y_true_values = np.asarray(list(map(lambda t:t[1],y_data)))
y_pred_values = np.asarray(list(map(lambda t:t[2],y_data)))
report = dskc_modeling.EvaluationReport(y_true_values, y_pred_values)
x.append(x_pos)
y.append(report)
x_pos += jump
return x,y
def test(model):
# read data
x_test, y_test, _ = util.read_test_data()
# predict
y_pred = model.predict(x_test)
# clean
array = []
for x in y_pred:
if x < 0:
array.append(0)
elif x > 1:
array.append(1)
else:
array.append(x)
y_pred = np.asarray(array)
# evaluate
report = dskc_modeling.EvaluationReport(y_test,
y_pred,
name="Bayesian Ridge")
return report
def test(model):
# read data
x_test, y_test, _ = util.read_test_data()
# predict
y_pred = model.predict(x_test)
# evaluate
report = dskc_modeling.EvaluationReport(y_test, y_pred, name="Logistic Regression")
return report
def test(model):
x_test, y_test, _ = util.read_test_data()
# predict
y_pred = model.predict(x_test)
# clean
y_pred = np.asarray([x[0] for x in y_pred])
# evaluate
report = dskc_modeling.EvaluationReport(y_test, y_pred, name="Neural Network")
return report
def test(model):
# read data
x_test, y_test, _ = util.read_test_data()
# predict
y_pred = model.predict(x_test)
# evaluate
report = dskc_modeling.EvaluationReport(y_test,
y_pred,
name="Random Forest Regressor",
to_binary=True)
return report