def gb():
exp(GradientBoostingRegressor,
shuffle=True,
randomSeeds=range(0, 100),
train_pctl=0.7,
train_vis=False,
save=True)
def lr():
exp(LinearRegression,
shuffle=True,
randomSeeds=range(0, 1),
train_pctl=0.7,
train_vis=True,
save=True)
def lasso():
exp(Lasso,
shuffle=True,
randomSeeds=range(6, 7),
train_pctl=0.7,
train_vis=True,
save=False,
alpha=0.015)
def ridge():
exp(Ridge,
shuffle=True,
randomSeeds=range(6, 7),
train_pctl=0.7,
train_vis=True,
save=False,
alpha=6)
def rf():
exp(RandomForestRegressor,
shuffle=True,
randomSeeds=range(0, 100),
train_pctl=0.7,
train_vis=False,
save=True,
n_estimators=100)
def knn():
exp(KNeighborsRegressor,
shuffle=True,
randomSeeds=range(0, 1),
train_pctl=0.7,
train_vis=True,
save=False,
n_neighbors=30,
weights="distance")
def __init__(self, operand, name, operation=None):
if operation is None:
self.distribution = operand.execute()
elif operation is "exp":
self.distribution = model.exp(operand.execute())
self.distribution.get_piecewise_pdf()
elif operation is "cos":
self.distribution = model.cos(operand.execute())
self.distribution.get_piecewise_pdf()
elif operation is "sin":
self.distribution = model.sin(operand.execute())
self.distribution.get_piecewise_pdf()
elif operation is "abs":
self.distribution = model.abs(operand)
self.distribution.get_piecewise_pdf()
else:
print("Unary operation not yet supported")
exit(-1)
self.operand = operand
self.name = name
self.operation = operation
self.sampleInit = True
self.a = self.distribution.range_()[0]
self.b = self.distribution.range_()[-1]
self.discretization = None
self.affine_error = None
self.symbolic_error = None
self.symbolic_affine = None
self.get_discretization()
self.constraints_dict = {}
self.collect_constraints()
def lasso_train():
train_r2 = []
val_r2 = []
aucs = []
alphas = [
1e-2, 1.25e-2, 1.5e-2, 1.75e-2, 2e-2, 2.25e-2, 2.5e-2, 2.75e-2, 3e-2,
3.5e-2, 4.0e-2, 5.0e-2, 1e-1, 5e-1, 1.0, 5, 10, 20, 50
]
for alpha in alphas:
train, val, auc = exp(Lasso,
shuffle=True,
randomSeeds=range(6, 7),
train_pctl=0.7,
train_vis=False,
save=True,
alpha=alpha)
train_r2.append(train)
val_r2.append(val)
aucs.append(auc)
res = {
"alpha": alphas,
"train": train_r2,
"validation": val_r2,
"auc": aucs
}
df = pd.DataFrame(res, columns=["alpha", "train", "validation", "auc"])
df.to_csv("exp_results_lasso.csv")
def ridge_train():
train_r2 = []
val_r2 = []
aucs = []
alphas = [
1e-10, 1e-5, 0.01, 0.1, 0.2, 0.5, 1, 2, 4, 6, 8, 10, 12, 14, 16, 32,
64, 128, 258, 512
]
for alpha in alphas:
train, val, auc = exp(Ridge,
shuffle=True,
randomSeeds=range(0, 100),
train_pctl=0.7,
train_vis=False,
save=True,
alpha=alpha)
train_r2.append(train)
val_r2.append(val)
aucs.append(auc)
res = {
"alpha": alphas,
"train": train_r2,
"validation": val_r2,
"auc": aucs
}
df = pd.DataFrame(res, columns=["alpha", "train", "validation", "auc"])
df.to_csv("exp_results_ridge.csv")