def lasso_lars(X_tr, y_tr, X_v, y_v, X_te, y_te, **kwargs):
'''
This function runs the lasso lars model on train, validate,
and test data with the option to include key word arguments
'''
# create lasso lars model
lars = LassoLars(**kwargs)
# fit the model to train data
lars.fit(X_tr, y_tr)
# fit the model to train data
lars_pred = lars.predict(X_tr)
# calculate the rmse on the train data
lars_rmse = sqrt(mean_squared_error(y_tr, lars_pred))
# predict the popularity on the validate data
lars_pred_v = lars.predict(X_v)
# calculate the rmse on the validate data
lars_rmse_v = sqrt(mean_squared_error(y_v, lars_pred_v))
# predict the popularity on the test data
lars_pred_t = lars.predict(X_te)
# calculate the rmse on the test data
lars_rmse_t = sqrt(mean_squared_error(y_te, lars_pred_t))
# print the train rmse
print('RMSE for LASSO + LARS \n')
print('On train data:\n', round(lars_rmse, 6), '\n')
return lars_rmse, lars_rmse_v, lars_rmse_t
def _lassolars(*,
train,
test,
x_predict=None,
metrics,
alpha=1.0,
fit_intercept=True,
verbose=False,
normalize=True,
precompute='auto',
max_iter=500,
eps=2.220446049250313e-16,
copy_X=True,
fit_path=True,
positive=False,
jitter=None,
random_state=None):
"""For more info visit :
https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LassoLars.html#sklearn.linear_model.LassoLars
"""
model = LassoLars(alpha=alpha,
fit_intercept=fit_intercept,
verbose=verbose,
normalize=normalize,
precompute=precompute,
max_iter=max_iter,
eps=eps,
copy_X=copy_X,
fit_path=fit_path,
positive=positive,
jitter=jitter,
random_state=random_state)
model.fit(train[0], train[1])
model_name = 'LassoLars'
y_hat = model.predict(test[0])
if metrics == 'mse':
accuracy = _mse(test[1], y_hat)
if metrics == 'rmse':
accuracy = _rmse(test[1], y_hat)
if metrics == 'mae':
accuracy = _mae(test[1], y_hat)
if x_predict is None:
return (model_name, accuracy, None)
y_predict = model.predict(x_predict)
return (model_name, accuracy, y_predict)
class LassoLarsPrim(primitive):
def __init__(self, random_state=0):
super(LassoLarsPrim, self).__init__(name='LassoLars')
self.hyperparams = []
self.type = 'Regressor'
self.description = "LassoLars is a lasso model implemented using the LARS algorithm, and unlike the implementation based on coordinate descent, this yields the exact solution, which is piecewise linear as a function of the norm of its coefficients."
self.hyperparams_run = {'default': True}
self.random_state = random_state
self.model = LassoLars(alpha=0.1)
self.accept_type = 'c_r'
def can_accept(self, data):
return self.can_accept_c(data, 'Regression')
def is_needed(self, data):
# data = handle_data(data)
return True
def fit(self, data):
data = handle_data(data)
self.model.fit(data['X'], data['Y'])
def produce(self, data):
output = handle_data(data)
output['predictions'] = self.model.predict(output['X'])
output['X'] = pd.DataFrame(output['predictions'], columns=[self.name+"Pred"])
final_output = {0: output}
return final_output
def LassoRegression(X_train, X_test, y_train, y_test):
regr = LassoLars(alpha=0.1)
print len(X_train.values.tolist()[0])
print len(X_train.values.tolist())
regr.fit(X_train.values.tolist(), y_train.values.tolist())
predictions = regr.predict(X_test)
return predictions
def Lasso(x_train, y_train, x_test, y_test):
estimator = LassoLars()
estimator.fit(x_train, y_train)
y_pred = estimator.predict(x_test)
mse_score = mse(y_test, y_pred)
print("mse_score: " + str(mse_score))
r2_score = r2(y_test, y_pred)
print("r2_score: " + str(r2_score))
def lasso_regression(args):
start = time.time()
with open(args.trainfile) as f:
train = np.genfromtxt(f, delimiter=',')
x = train[:, :-1]
fe = FeatureEngineering(30, 2, 400)
x = fe.fit_transform(x)
x = np.column_stack((np.ones(x.shape[0], ), x))
y = train[:, -1]
kf = KFold(n_splits=10)
kf.get_n_splits(x)
ls = [0.003]
min_error = float('inf')
min_l = None
for l in ls:
error_sum = 0
for train_index, test_index in kf.split(x):
x_train, x_test = x[train_index], x[test_index]
y_train, y_test = y[train_index], y[test_index]
reg = LassoLars(alpha=l)
reg.fit(x_train, y_train)
error = (np.linalg.norm(reg.predict(x_test) - y_test)**
2) / (2 * x_test.shape[0])
error_sum += error
if error_sum < min_error:
min_error = error_sum
min_l = l
reg = LassoLars(alpha=min_l)
reg.fit(x, y)
w = reg.coef_
error = (np.linalg.norm(reg.predict(x) - y)**2) / (2 * x.shape[0])
print('Lambda: ', min_l, '. Error: ', error)
with open(args.testfile) as f:
test = np.genfromtxt(f, delimiter=',')
x = fe.transform(test)
x = np.column_stack((np.ones(test.shape[0], ), x))
predictions = x @ w
np.savetxt(args.outputfile, predictions)
print('Time: ', time.time() - start)
def predict_LarsLasso(X, y, train, test, alpha=0.1):
# Fit
lars = LassoLars(alpha)
lars.fit(X.iloc[train], y.iloc[train])
# Predict
prediction = lars.predict(X.iloc[test])
return prediction
def LassoLarsTest(dataMat, labelMat):
clf1 = LassoLars(alpha=1, max_iter=100)
clf1.fit(dataMat[0:99], labelMat[0:99])
labelTest1 = clf1.predict(dataMat[100:199])
print('LassoLars ', ((labelTest1 - labelMat[100:199])**2).sum())
clf2 = LassoLarsCV(max_n_alphas=10, max_iter=100)
clf2.fit(dataMat[0:99], labelMat[0:99])
labelTest2 = clf2.predict(dataMat[100:199])
print('LassoLarsCV', ((labelTest2 - labelMat[100:199])**2).sum())
def lasso_lars(X, y):
#train model
lars = LassoLars(alpha=0.1)\
.fit(X, y)
lars_pred = lars.predict(X)
lars_rmse = sqrt(mean_squared_error(y, lars_pred))
return lars_rmse
def ll_validate_test(X, y, X_vt, y_vt):
#train model
lars = LassoLars(alpha=0.1)\
.fit(X, y)
#validate model
lars_pred_v = lars.predict(X_vt)
lars_rmse_v = sqrt(mean_squared_error(y_vt, lars_pred_v))
return lars_rmse_v
class in_lassoLars(regression):
def trainAlgo(self):
self.model = LassoLars(alpha=self.param['alpha'],
normalize=self.param['normalize'],
fit_intercept=self.param['fit_intercept'],
max_iter=self.param['max_iter'],
positive=self.param['positive'])
self.model.fit(self.inputData['X'], self.outputData['Y'])
def predictAlgo(self):
self.result['Y'] = self.model.predict(self.inputData['X'])
def lasso_lars_test(x_scaleddf, target, X_test, y_test):
'''
runs Lasso Lars algorithm
'''
# Make a model
lars = LassoLars(alpha=1)
# Fit a model
lars.fit(x_scaleddf, target)
# Make Predictions
lars_pred = lars.predict(X_test)
# calculate MAE
lars_MAE = mean_absolute_error(y_test, lars_pred)
return lars_MAE, lars, lars_pred
def fit_model_11(self,toWrite=False):
model = LassoLars(alpha=1,max_iter=5000)
for data in self.cv_data:
X_train, X_test, Y_train, Y_test = data
model.fit(X_train,Y_train)
pred = model.predict(X_test)
print("Model 11 score %f" % (logloss(Y_test,pred),))
if toWrite:
f2 = open('model11/model.pkl','w')
pickle.dump(model,f2)
f2.close()
def lasso_lars(x_scaleddf, target):
'''
runs Lasso Lars algorithm
'''
# Make a model
lars = LassoLars(alpha=1)
# Fit a model
lars.fit(x_scaleddf, target)
# Make Predictions
lars_pred = lars.predict(x_scaleddf)
# Computer root mean squared error
lars_rmse = sqrt(mean_squared_error(target, lars_pred))
return lars_rmse
class _LassoLarsImpl:
def __init__(self, **hyperparams):
self._hyperparams = hyperparams
self._wrapped_model = Op(**self._hyperparams)
def fit(self, X, y=None):
if y is not None:
self._wrapped_model.fit(X, y)
else:
self._wrapped_model.fit(X)
return self
def predict(self, X):
return self._wrapped_model.predict(X)
class DkuLassoLarsRegressor(BaseEstimator):
def __init__(self, max_var=0):
self.max_var = max_var
self.lars = None
self.X_offset = None
self.y_offset = None
self.X_scale = None
self.coef_ = None
self.current_index = None
self.intercept_ = None
self.coef_path_ = None
def fit(self, X, y):
# note: for now we perform rescaling. While this requires some more computation on our part, it has better
# numerical stability (could test with or without)
self.lars = LassoLars(alpha=0.0).fit(X, y)
# we recreate the rescaling
_, _, self.X_offset, self.y_offset, self.X_scale = self.lars._preprocess_data(
X, y, True, True, True)
# we normalize the coef path here
self.coef_path_ = [x / self.X_scale for x in self.lars.coef_path_.T]
self.coef_ = self.lars.coef_
self.intercept_ = self.lars.intercept_
self.alphas = self.lars.alphas_
if self.max_var > 0:
self._perform_cut(self.max_var)
return self
def _perform_cut(self, n):
n = min(n, self.lars.coef_path_.shape[1] - 1)
self.current_index = n
# note: not normalized, this is normal since the _set_intercept will normalize it
coef = self.lars.coef_path_[:, n]
self.lars.coef_ = coef
# recompute the intercept and normalize coefficients using scikit private method
self.lars._set_intercept(self.X_offset, self.y_offset, self.X_scale)
self.coef_ = self.lars.coef_
def post_process(self, user_meta):
if self.current_index is not None:
n = self.current_index
else:
n = self.max_var
n = user_meta.get("lars_cut", n)
if n > 0:
self._perform_cut(n)
def predict(self, X):
return self.lars.predict(X)
def linear_regressor(x, target, causes):
""" Regression and prediction using a lasso
:param x: data
:param target: target - effect
:param causes: causes of the causal mechanism
:return: regenerated data with the fitted model
"""
if len(causes) == 0:
x = np.random.normal(size=(target.shape[0], 1))
lasso = LassoLars(alpha=1.) # no regularization
lasso.fit(x, target)
return lasso.predict(x)
def KFoldValidationLasso(X, Y, lam, k=10):
loss = 0
for i in range(k):
start = math.floor(X.shape[0] * i / k)
end = math.floor(X.shape[0] * (i + 1) / k)
x = np.r_[X[:start, :], X[end:, :]]
y = np.r_[Y[:start], Y[end:]]
vx = X[start:end, :]
vy = Y[start:end]
model = LassoLars(alpha=lam)
model.fit(x, y)
vyhat = model.predict(vx)
loss += (np.linalg.norm(vy - vyhat, ord=2) /
np.linalg.norm(vy, ord=2))**2
loss /= k
return loss
def test_simple_vs_refined_algorithm(theta, fit_path):
# Test the consistency of the results between the 2 versions of
# the algorithm.
# Simple Algorithm (2 steps of Lasso Lars)
lasso1 = LassoLars(alpha=alpha)
lasso1.fit(X_train, y_train)
X1 = X_train.copy()
X1[:, lasso1.coef_ == 0] = 0
lasso2 = LassoLars(alpha=alpha*theta)
lasso2.fit(X1, y_train)
pred_simple = lasso2.predict(X_test)
# Refined Algorithm
relasso = RelaxedLassoLars(alpha=alpha, theta=theta, fit_path=fit_path)
relasso.fit(X_train, y_train)
pred_refined = relasso.predict(X_test)
assert_array_almost_equal(pred_simple, pred_refined)
assert_array_almost_equal(lasso2.coef_, relasso.coef_)
assert_almost_equal(lasso2.score(X_test, y_test),
relasso.score(X_test, y_test),
decimal=2)
y_split = np.array_split(y,10)
from sklearn.linear_model import LassoLars
l = [0.0,0.001,0.002,0.005,0.01,0.02,0.05,0.1,0.2,0.5,1,2,5,10,20,50,100]
nmse = np.zeros(len(l))
for i in range(len(l)):
alpha = l[i]
regressor = LassoLars(alpha)
for j in range(10):
###
temp = list(range(10))
temp.remove(j)
Xi_test = X_split[j]
Xi = np.concatenate([X_split[k] for k in temp])
yi_test = y_split[j]
yi = np.concatenate([y_split[k] for k in temp])
###
regressor.fit(Xi, yi)
###
yi_pred = regressor.predict(Xi_test)
nmse[i]+=compute_error(yi_test,yi_pred)
###
nmse = nmse/10
###
alpha = l[np.argmin(nmse)]
regressor = LassoLars(alpha)
regressor.fit(X, y)
y_pred = regressor.predict(X_test)
f = open(sys.argv[4], "a")
for i in range(len(y_pred)):
f.write(str(y_pred[i])+'\n')
f.close()
def lassoLars(X, y, value):
regressor = LassoLars(alpha=0.3, max_iter=600000)
regressor.fit(X, y)
y_pred = regressor.predict(value)
return y_pred
# LassoLars Regression # The Least Angle Regression (LARS) can be used as an alternative method for calculating Least Absolute Shrinkage # and Selection Operator (LASSO) fit. import numpy as np from sklearn import datasets from sklearn.linear_model import LassoLars # load the iris datasets dataset = datasets.load_diabetes() # fit a LASSO using LARS model to the data model = LassoLars(alpha=0.1) model.fit(dataset.data, dataset.target) print(model) # make predictions expected = dataset.target predicted = model.predict(dataset.data) # summarize the fit of the model mse = np.mean((predicted - expected)**2) print(mse) print(model.score(dataset.data, dataset.target))
def task2(data):
df = data
dfreg = df.loc[:, ['Adj Close', 'Volume']]
dfreg['HL_PCT'] = (df['High'] - df['Low']) / df['Close'] * 100.0
dfreg['PCT_change'] = (df['Close'] - df['Open']) / df['Open'] * 100.0
# Drop missing value
dfreg.fillna(value=-99999, inplace=True)
# We want to separate 1 percent of the data to forecast
forecast_out = int(math.ceil(0.01 * len(dfreg)))
# Separating the label here, we want to predict the AdjClose
forecast_col = 'Adj Close'
dfreg['label'] = dfreg[forecast_col].shift(-forecast_out)
X = np.array(dfreg.drop(['label'], 1))
# Scale the X so that everyone can have the same distribution for linear regression
X = preprocessing.scale(X)
# Finally We want to find Data Series of late X and early X (train) for model generation and evaluation
X_lately = X[-forecast_out:]
X = X[:-forecast_out]
# Separate label and identify it as y
y = np.array(dfreg['label'])
y = y[:-forecast_out]
#Split data
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=0)
##################
##################
##################
# Linear regression
clfreg = LinearRegression(n_jobs=-1)
clfreg.fit(X_train, y_train)
# Quadratic Regression 2
clfpoly2 = make_pipeline(PolynomialFeatures(2), Ridge())
clfpoly2.fit(X_train, y_train)
# Quadratic Regression 3
clfpoly3 = make_pipeline(PolynomialFeatures(3), Ridge())
clfpoly3.fit(X_train, y_train)
# KNN Regression
clfknn = KNeighborsRegressor(n_neighbors=2)
clfknn.fit(X_train, y_train)
# Lasso Regression
clflas = Lasso()
clflas.fit(X_train, y_train)
# Multitask Lasso Regression
# clfmtl = MultiTaskLasso(alpha=1.)
# clfmtl.fit(X_train, y_train).coef_
# Bayesian Ridge Regression
clfbyr = BayesianRidge()
clfbyr.fit(X_train, y_train)
# Lasso LARS Regression
clflar = LassoLars(alpha=.1)
clflar.fit(X_train, y_train)
# Orthogonal Matching Pursuit Regression
clfomp = OrthogonalMatchingPursuit(n_nonzero_coefs=2)
clfomp.fit(X_train, y_train)
# Automatic Relevance Determination Regression
clfard = ARDRegression(compute_score=True)
clfard.fit(X_train, y_train)
# Logistic Regression
# clflgr = linear_model.LogisticRegression(penalty='l1', solver='saga', tol=1e-6, max_iter=int(1e6), warm_start=True)
# coefs_ = []
# for c in cs:
# clflgr.set_params(C=c)
# clflgr.fit(X_train, y_train)
# coefs_.append(clflgr.coef_.ravel().copy())
clfsgd = SGDRegressor(random_state=0, max_iter=1000, tol=1e-3)
clfsgd.fit(X_train, y_train)
##################
##################
##################
#Create confindence scores
confidencereg = clfreg.score(X_test, y_test)
confidencepoly2 = clfpoly2.score(X_test, y_test)
confidencepoly3 = clfpoly3.score(X_test, y_test)
confidenceknn = clfknn.score(X_test, y_test)
confidencelas = clflas.score(X_test, y_test)
# confidencemtl = clfmtl.score(X_test, y_test)
confidencebyr = clfbyr.score(X_test, y_test)
confidencelar = clflar.score(X_test, y_test)
confidenceomp = clfomp.score(X_test, y_test)
confidenceard = clfard.score(X_test, y_test)
confidencesgd = clfsgd.score(X_test, y_test)
# results
print('The linear regression confidence is:', confidencereg * 100)
print('The quadratic regression 2 confidence is:', confidencepoly2 * 100)
print('The quadratic regression 3 confidence is:', confidencepoly3 * 100)
print('The knn regression confidence is:', confidenceknn * 100)
print('The lasso regression confidence is:', confidencelas * 100)
# print('The lasso regression confidence is:',confidencemtl*100)
print('The Bayesian Ridge regression confidence is:', confidencebyr * 100)
print('The Lasso LARS regression confidence is:', confidencelar * 100)
print('The OMP regression confidence is:', confidenceomp * 100)
print('The ARD regression confidence is:', confidenceard * 100)
print('The SGD regression confidence is:', confidencesgd * 100)
#Create new columns
forecast_reg = clfreg.predict(X_lately)
forecast_pol2 = clfpoly2.predict(X_lately)
forecast_pol3 = clfpoly3.predict(X_lately)
forecast_knn = clfknn.predict(X_lately)
forecast_las = clflas.predict(X_lately)
forecast_byr = clfbyr.predict(X_lately)
forecast_lar = clflar.predict(X_lately)
forecast_omp = clfomp.predict(X_lately)
forecast_ard = clfard.predict(X_lately)
forecast_sgd = clfsgd.predict(X_lately)
#Process all new columns data
dfreg['Forecast_reg'] = np.nan
last_date = dfreg.iloc[-1].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_reg:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg.loc[next_date] = [np.nan for _ in range(len(dfreg.columns))]
dfreg['Forecast_reg'].loc[next_date] = i
dfreg['Forecast_pol2'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_pol2:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_pol2'].loc[next_date] = i
dfreg['Forecast_pol3'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_pol3:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_pol3'].loc[next_date] = i
dfreg['Forecast_knn'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_knn:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_knn'].loc[next_date] = i
dfreg['Forecast_las'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_las:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_las'].loc[next_date] = i
dfreg['Forecast_byr'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_byr:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_byr'].loc[next_date] = i
dfreg['Forecast_lar'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_lar:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_lar'].loc[next_date] = i
dfreg['Forecast_omp'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_omp:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_omp'].loc[next_date] = i
dfreg['Forecast_ard'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_ard:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_ard'].loc[next_date] = i
dfreg['Forecast_sgd'] = np.nan
last_date = dfreg.iloc[-26].name
last_unix = last_date
next_unix = last_unix + datetime.timedelta(days=1)
for i in forecast_sgd:
next_date = next_unix
next_unix += datetime.timedelta(days=1)
dfreg['Forecast_sgd'].loc[next_date] = i
return dfreg.index.format(formatter=lambda x: x.strftime(
'%Y-%m-%d')), dfreg['Adj Close'].to_list(
), dfreg['Forecast_reg'].to_list(), dfreg['Forecast_pol2'].to_list(
), dfreg['Forecast_pol3'].to_list(), dfreg['Forecast_knn'].to_list(
), dfreg['Forecast_las'].to_list(), dfreg['Forecast_byr'].to_list(
), dfreg['Forecast_lar'].to_list(), dfreg['Forecast_omp'].to_list(
), dfreg['Forecast_ard'].to_list(), dfreg['Forecast_sgd'].to_list()
rank_result['Lars_pca'] = sumsum / float(result_row)
rs_score['Lars_pca'] = r2_score(y_test, y)
LarsModel = Lars()
LarsModel.fit(X_train_std, y_train)
y = LarsModel.predict(X_test_std)
[result_row] = y.shape
sumsum = 0
#print y
for i in range(result_row):
sumsum = sumsum + (y[i] - y_test[i]) * (y[i] - y_test[i])
rank_result['Lars_std'] = sumsum / float(result_row)
rs_score['Lars_std'] = r2_score(y_test, y)
LassoLarsModel = LassoLars()
LassoLarsModel.fit(X_train_pca, y_train)
y = LassoLarsModel.predict(X_test_pca)
[result_row] = y.shape
sumsum = 0
#print y
for i in range(result_row):
sumsum = sumsum + (y[i] - y_test[i]) * (y[i] - y_test[i])
rank_result['LassoLars_pca'] = sumsum / float(result_row)
rs_score['LassoLars_pca'] = r2_score(y_test, y)
LassoLarsModel = LassoLars()
LassoLarsModel.fit(X_train_std, y_train)
y = LassoLarsModel.predict(X_test_std)
[result_row] = y.shape
sumsum = 0
#print y
for i in range(result_row):
sumsum = sumsum + (y[i] - y_test[i]) * (y[i] - y_test[i])
def all_models_info():
'''takes in data
sets baseline
sets SSE, MSE, and RMSE
returns infor for all 4'''
# get data
df = acquire.acquire_zillow()
df = prepare.clean_zillow(df)
df = prepare.focused_zillow(df)
# pull from add to trian
train = evaluate.add_to_train()
X_train, y_train, X_validate, y_validate, X_test, y_test = evaluate.xtrain_xval_xtest(
)
#OLS Model
lm = LinearRegression(normalize=True)
lm.fit(X_train, y_train.appraised_value)
y_train['appraised_value_pred_lm'] = lm.predict(X_train)
rmse_train_lm = mean_squared_error(
y_train.appraised_value, y_train.appraised_value_pred_lm)**(1 / 2)
y_validate['appraised_value_pred_lm'] = lm.predict(X_validate)
rmse_validate_lm = mean_squared_error(
y_validate.appraised_value,
y_validate.appraised_value_pred_lm)**(1 / 2)
#LARS Model
lars = LassoLars(alpha=1.0)
lars.fit(X_train, y_train.appraised_value)
y_train['appraised_value_pred_lars'] = lars.predict(X_train)
rmse_train_lars = mean_squared_error(
y_train.appraised_value, y_train.appraised_value_pred_lars)**1 / 2
y_validate['appraised_value_pred_lars'] = lars.predict(X_validate)
rmse_validate_lars = mean_squared_error(
y_validate.appraised_value,
y_validate.appraised_value_pred_lars)**1 / 2
#GLM
glm = TweedieRegressor(power=1, alpha=0)
glm.fit(X_train, y_train.appraised_value)
y_train['appraised_value_pred_glm'] = glm.predict(X_train)
rmse_train_glm = mean_squared_error(
y_train.appraised_value, y_train.appraised_value_pred_glm)**1 / 2
y_validate['appraised_value_pred_glm'] = glm.predict(X_validate)
rmse_validate_glm = mean_squared_error(
y_validate.appraised_value, y_validate.appraised_value_pred_glm)**1 / 2
# PF
pf = PolynomialFeatures(degree=2)
X_train_degree2 = pf.fit_transform(X_train)
X_validate_degree2 = pf.transform(X_validate)
X_test_degree2 = pf.transform(X_test)
# LM2
lm2 = LinearRegression(normalize=True)
lm2.fit(X_train_degree2, y_train.appraised_value)
y_train['appraised_value_pred_lm2'] = lm2.predict(X_train_degree2)
rmse_train_lm2 = mean_squared_error(
y_train.appraised_value, y_train.appraised_value_pred_lm2)**1 / 2
y_validate['appraised_value_pred_lm2'] = lm2.predict(X_validate_degree2)
rmse_validate_lm2 = mean_squared_error(
y_validate.appraised_value, y_validate.appraised_value_pred_lm2)**1 / 2
print("RMSE for OLS using LinearRegression\nTraining/In-Sample: ",
rmse_train_lm, "\nValidation/Out-of-Sample: ", rmse_validate_lm)
print("--------------------------------------------------------------")
print("RMSE for Lasso + Lars\nTraining/In-Sample: ", rmse_train_lars,
"\nValidation/Out-of-Sample: ", rmse_validate_lars)
print("--------------------------------------------------------------")
print(
"RMSE for GLM using Tweedie, power=1 & alpha=0\nTraining/In-Sample: ",
rmse_train_glm, "\nValidation/Out-of-Sample: ", rmse_validate_glm)
print("--------------------------------------------------------------")
print("RMSE for Polynomial Model, degrees=2\nTraining/In-Sample: ",
rmse_train_lm2, "\nValidation/Out-of-Sample: ", rmse_validate_lm2)
lars_alpha, lars_err
##
max_iter = 50000
lasso_model = Lasso(alpha=alpha_list[0], max_iter=max_iter).fit(trainX, trainY)
elasticNet_model = ElasticNet(alpha=alpha_list[1],
max_iter=max_iter).fit(trainX, trainY)
ridge_model = Ridge(alpha=alpha_list[2], max_iter=max_iter).fit(trainX, trainY)
lars_model = LassoLars(alpha=alpha_list[3],
max_iter=max_iter).fit(trainX, trainY)
lasso_pred = np.expm1(lasso_model.predict(raw_test_df))
ridge_pred = np.expm1(ridge_model.predict(raw_test_df))
elasticNet_pred = np.expm1(elasticNet_model.predict(raw_test_df))
lars_pred = np.expm1(lars_model.predict(raw_test_df))
pred_list = np.array(
[lasso_pred, ridge_pred, elasticNet_pred, lars_pred, xgb_pred])
# take average of 4 models
err_list.append(xgb_err)
err_list = np.array(err_list)
w_list = 1 / err_list
total_w = np.sum(w_list)
predictions = np.matmul(w_list / total_w, pred_list)
# xgb_w, lasso_w, elas_w, ridge_w, lars_w = 1/xgb_err, 1/lasso_err, 1/elas_err, 1/ridge_err, 1/lars_err
# total_w = xgb_w + lasso_w + elas_w + ridge_w + lars_w
# predictions = lasso_w/total_w*lasso_pred + ridge_w/total_w*ridge_pred + \
# elas_w/total_w*elasticNet_pred + xgb_w/total_w*xgb_pred +
# print (MSELasso(y_test,pred.reshape((pred.size,1))))
vals = [0.0000001, 0.0001, 1, 10]
errors = np.empty(4)
for j in range(4):
lm = vals[j]
k = 4
err = np.empty(k)
l = int(np.ma.size(x_train, axis=0) / k)
x_cv, x_tr = np.split(x_train.copy(), [l], axis=0)
y_cv, y_tr = np.split(y_train.copy(), [l], axis=0)
model = LassoLars(alpha=lm)
model.fit(x_tr, y_tr.ravel())
pred = model.predict(x_cv)
err[0] = MSELasso(y_cv, pred.reshape((pred.size, 1)))
for i in range(k - 1):
x_tr[i * l:(i + 1) * l], x_cv = x_cv, x_tr[i * l:(i + 1) *
l].copy()
y_tr[i * l:(i + 1) * l], y_cv = y_cv, y_tr[i * l:(i + 1) *
l].copy()
model = LassoLars(alpha=lm)
model.fit(x_tr, y_tr.ravel())
pred = model.predict(x_cv)
err[i + 1] = MSELasso(y_cv, pred.reshape((pred.size, 1)))
errors[j] = np.mean(err)
x_tr = np.concatenate((x_train, np.square(x_train), np.power(x_train, 3)),
# Linear Regression
linear_reg = LinearRegression()
linear_reg.fit(X_train, Y_train)
Y_pred = linear_reg.predict(X_test)
linear_r2 = r2_score(Y_expected, Y_pred)
linear_mse = mean_squared_error(Y_expected, Y_pred)
print("Linear Regression\n", "R2: ", linear_r2, "MSE:", linear_mse)
plot_prediction("Linear Regression", Y_pred, test['close'])
# Lasso Lars
lassolars_reg = LassoLars()
lassolars_reg.fit(X_train, Y_train)
Y_pred = lassolars_reg.predict(X_test)
lassolars_r2 = r2_score(Y_expected, Y_pred)
lassolars_mse = mean_squared_error(Y_expected, Y_pred)
print("Lasso Lars Regression\n", "R2: ", lassolars_r2, "MSE:", lassolars_mse)
plot_prediction("Lasso Lars Regression", Y_pred, test['close'])
# Theil Sen Regressor
theil_reg = TheilSenRegressor()
theil_reg.fit(X_train, Y_train)
Y_pred = theil_reg.predict(X_test)
theil_r2 = r2_score(Y_expected, Y_pred)
theil_mse = mean_squared_error(Y_expected, Y_pred)
print("Theil Sen Regression\n", "R2: ", theil_r2, "MSE:", theil_mse)
plot_prediction("Theil Sen Regression", Y_pred, test['close'])
train = pd.read_csv("train/subtrain.csv", chunksize=100000, iterator=True)
all_classes = np.array([0, 1])
for chunk in train:
y_train = chunk["click"]
chunk = chunk[cols]
chunk = chunk.join(
pd.DataFrame([dayhour(x) for x in chunk.hour], columns=["wd", "hr"]))
chunk.drop(["hour"], axis=1, inplace=True)
Xcat = fh.transform(np.asarray(chunk.astype(str)))
clf.fit(Xcat, y_train)
# Create a submission file
usecols = cols + ["id"]
X_test = pd.read_csv("test/mtest.csv", usecols=usecols)
X_test = X_test.join(
pd.DataFrame([dayhour(x) for x in X_test.hour], columns=["wd", "hr"]))
X_test.drop(["hour"], axis=1, inplace=True)
X_enc_test = fh.transform(np.asarray(X_test.astype(str)))
y_act = pd.read_csv("test/mtest.csv", usecols=['click'])
y_pred = clf.predict(X_enc_test)
with open('logloss.txt', 'a') as f:
f.write('\n' + str(log_loss(y_act, y_pred)))
with open("submission/submission_elnet.csv", "w") as f:
f.write("id,click\n")
for idx, xid in enumerate(X_test.id):
f.write(str(xid) + "," + "{0:.10f}".format(y_pred[idx]) + "\n")
f.close()
from sklearn import linear_model
lassy = linear_model.Lasso(alpha=0.001)
lassy.fit(X_train_scaled, y_train)
y_pred = lassy.predict(X_test_scaled)
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
# In[19]:
laslars = LassoLars(alpha=0.0001)
laslars.fit(X_train_scaled, y_train)
y_pred = laslars.predict(X_test_scaled)
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
# In[20]:
from sklearn.linear_model import ElasticNet
elastic = ElasticNet(random_state=0)
elastic.fit(X_train_scaled, y_train)
y_pred = elastic.predict(X_test_scaled)
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
def dt_process(df2,option_slctd):
df = df2.copy() #work with a local copy
opted_country = option_slctd # 'Brazil' # input("Select the country - ")
print(opted_country)
dt_one_country = df[df["location"] == opted_country][['date', 'new_cases']] #work the predictions only for the column 'new_cases' in the rest of code
dt_one_country['new_cases'] = dt_one_country['new_cases'].fillna(0)
dt_one_country['date'] = pd.to_datetime(dt_one_country['date'])
dt_one_country['Days Since'] = dt_one_country['date'] - dt_one_country['date'].min()
dt_one_country['Days Since'] = dt_one_country['Days Since'].dt.days #use the days since the starting date of records of this country, use this as the known variable to make the prediction
train_ml = dt_one_country.iloc[:int(dt_one_country.shape[0] * 0.95)] #First 95% dates used for fitting the regressor
valid_ml = dt_one_country.iloc[int(dt_one_country.shape[0] * 0.95):] #last 5% dates to be predicted and compared to validation data of these dates
fitinput_x = np.array(train_ml["Days Since"]).reshape(-1, 1) #data should be in arrays for regressors, i think, have to cross check this Days Since is the known x data
fitinput_y = np.array(train_ml["new_cases"]).reshape(-1, 1) # new_cases is the y data, this data is used to 'fit' the regressor
# linreg = LinearRegression(normalize=True) #use this Linear Regressor model 'lin_reg' to fit and predict
Larspd = LassoLars(alpha=.1)
Larspd.fit(fitinput_x, fitinput_y) #fitting the regressor
x_pred = np.array(valid_ml["Days Since"]).reshape(-1, 1)
y_pred = Larspd.predict(x_pred) #predicting using regressor for the 5% days
model_scores = []
model_scores.append(np.sqrt(mean_squared_error(valid_ml["new_cases"], y_pred)))
# lin_reg.score(x_pred,valid_ml['new_cases'])
# print(np.sqrt(mean_squared_error(valid_ml["new_cases"], y_pred)))
# plt.figure(figsize=(11, 6))
prediction_linreg = Larspd.predict(np.array(dt_one_country["Days Since"]).reshape(-1, 1)) #use this as predictor for all the days, to understand the fitting line
linreg_output = []
# print("i am predicting ")
for i in range(prediction_linreg.shape[0]):
linreg_output.append(prediction_linreg[i])#[0])
# print("i am before figure ")
fig_LarsReg = go.Figure() #this handle can be returned to plot the figure outside of this function
#not currently returned
#shows the original recorded data for all the days
fig_LarsReg.add_trace(go.Scatter(x=dt_one_country['date'], y=dt_one_country["new_cases"],
mode='lines+markers', name="Train Data for new Cases"))
#shows the predicted data for all the days
fig_LarsReg.add_trace(go.Scatter(x=valid_ml['date'], y=y_pred,
mode='lines', name="Lars Regression Best Fit Line",
line=dict(color='red', dash='dot')))
# fig_LarsReg.add_trace(go.Scatter(x=dt_one_country['date'], y=linreg_output,
# mode='lines', name="Linear Regression Best Fit Line",
# line=dict(color='black', dash='dot')))
fig_LarsReg.add_vline(x=valid_ml['date'].iloc[0], line_dash="dash") # ,#add vertical line on the date to know the SPLIT between training and test data
fig_LarsReg.update_layout(title="new Cases Lars Regression Prediction " + str(opted_country),
xaxis_title="Date", yaxis_title="new Cases", legend=dict(x=0, y=1, traceorder="normal"))
# fig_LarsReg.show()
poly = PolynomialFeatures(degree=8) #Polynomial regressor initiate the model
train_poly = poly.fit_transform(fitinput_x) #do not know why we need this fit_transform specifically for Polynomial method
fitin_valid = np.array(valid_ml["Days Since"]).reshape(-1, 1)
valid_poly = poly.fit_transform(fitin_valid)
y_train_to_compare = train_ml['new_cases']
lin_reg = LinearRegression(normalize=True)
lin_reg.fit(train_poly, y_train_to_compare)
prediction_poly = lin_reg.predict(valid_poly)
lin_reg.score(valid_poly, valid_ml['new_cases'].values)
# print(np.sqrt(mean_squared_error(valid_ml["new_cases"], prediction_poly)))
model_scores.append(np.sqrt(mean_squared_error(valid_ml["new_cases"], prediction_poly))) #use this score to compare predictors and to know how close the predicted data is with the real known data
additional_30days = np.linspace(1, 30, 30) #predict additionally for 30days not in record, to know how the curve progresses
pred_input_compiled_data = []
pred_input_compiled_data = np.array(dt_one_country["Days Since"]).reshape(-1, 1)
pred_input_compiled_data = np.append(pred_input_compiled_data, pred_input_compiled_data[-1] + additional_30days)
# add_pred_dates = pd.DataFrame(columns=['date'])
add_pred_dates = dt_one_country['date']
for i in range(1, 31):
add_pred_dates = add_pred_dates.append(add_pred_dates.iloc[-1:] + timedelta(days=1), ignore_index=True) #increment the days count for the 30added days using datetime class
# comp_data=poly.fit_transform(np.array(dt_one_country["Days Since"]).reshape(-1,1))
comp_data = poly.fit_transform(pred_input_compiled_data.reshape(-1, 1))
# plt.figure(figsize=(11, 6))
predictions_poly = lin_reg.predict(comp_data)
fig_PolyReg = go.Figure() #returning this handle to show figure outside the function
fig_PolyReg.add_trace(go.Scatter(x=dt_one_country['date'], y=dt_one_country["new_cases"],
mode='lines+markers', name="Train Data for new Cases in " + str(opted_country)))
# fig.add_trace(go.Scatter(x=dt_one_country['date'], y=predictions_poly,
fig_PolyReg.add_trace(go.Scatter(x=add_pred_dates, y=predictions_poly,
mode='lines', name="Polynomial Regression Best Fit",
line=dict(color='red', dash='dot')))
fig_PolyReg.add_vline(x=valid_ml['date'].iloc[0], line_dash="dash") # ,#add vertical line on the date to know the SPLIT between training and test data
fig_PolyReg.update_layout(title="new Cases Polynomial Regression Prediction",
xaxis_title="Date", yaxis_title="new Cases",
legend=dict(x=0, y=1, traceorder="normal"))
# fig_PolyReg.show()
# train_ml=dt_one_country.iloc[:int(dt_one_country.shape[0]*0.95)]
# valid_ml=dt_one_country.iloc[int(dt_one_country.shape[0]*0.95):]
model_train = dt_one_country.iloc[:int(dt_one_country.shape[0] * 0.95)]
valid = dt_one_country.iloc[int(dt_one_country.shape[0] * 0.95):]
y_pred = valid.copy()
#there is no x,y data for fitting using Holts model --- just pass the known data, that is new_cases for the known days
holt = Holt(np.asarray(model_train["new_cases"])).fit(smoothing_level=0.9, smoothing_trend=0.4, optimized=False) #Holt model, smoothing parameters can be varied to observe behavior
y_pred["Holt"] = holt.forecast(len(valid)) #how many data to predict
# y_holt_pred["Holt"]=holt.forecast(len(valid)+30)
# print(np.sqrt(mean_squared_error(y_pred["new_cases"], y_pred["Holt"])))
model_scores.append(np.sqrt(mean_squared_error(y_pred["new_cases"], y_pred["Holt"])))
fig_Holt = go.Figure()
fig_Holt.add_trace(go.Scatter(x=model_train['date'], y=model_train["new_cases"],
mode='lines+markers', name="Train Data for new Cases " + str(opted_country)))
fig_Holt.add_trace(go.Scatter(x=valid['date'], y=valid["new_cases"],
mode='lines+markers', name="Validation Data for new Cases " + str(opted_country)))
fig_Holt.add_vline(x=valid['date'].iloc[0], line_dash="dash") # ,#add vertical line on the date to know the SPLIT between training and test data
fig_Holt.add_trace(go.Scatter(x=valid['date'], y=y_pred["Holt"],
mode='lines+markers', name="Prediction of new Cases " + str(opted_country)))
fig_Holt.update_layout(title="new Cases Holt's Linear Model Prediction",
xaxis_title="Date", yaxis_title="new Cases", legend=dict(x=0, y=1, traceorder="normal"))
# fig_Holt.show()
# the following is Log Linear predictor not currently shown in figure
x_train = train_ml['Days Since']
y_train_1 = train_ml['new_cases']
y_train_1 = y_train_1.astype('float64')
y_train_1 = y_train_1.apply(lambda x: np.log1p(x)) #first take logarithm of data and then use Linear predictor
y_train_1.replace([np.inf, -np.inf], 0, inplace=True)
x_test = valid_ml['Days Since']
y_test = valid_ml['new_cases']
# y_test = y_test.astype('float64')
# y_test = y_test.apply(lambda x: np.log1p(x))
# y_test.replace([np.inf, -np.inf], 0, inplace=True)
regr = LinearRegression(normalize=True)
regr.fit(np.array(x_train).reshape(-1, 1), np.array(y_train_1).reshape(-1, 1))
ypred = regr.predict(np.array(x_test).reshape(-1, 1))
# print(np.sqrt(mean_squared_error(y_test, np.expm1(ypred))))
# # Plot results
# fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(15, 6))
#
# ax1.plot(valid_ml['date'], np.expm1(ypred))
# ax1.plot(dt_one_country['date'], dt_one_country['new_cases'])
# ax1.axvline(valid_ml['date'].iloc[0], linewidth=2, ls=':', color='grey', alpha=0.5)
# ax1.legend(['Predicted cases', 'Actual cases', 'Train-test split'], loc='upper left')
# ax1.set_xlabel("Day count ")
# ax1.set_ylabel("new Cases")
#
# ax2.plot(valid_ml['date'], ypred)
# ax2.plot(dt_one_country['date'], np.log1p(dt_one_country['new_cases']))
# ax2.axvline(valid_ml['date'].iloc[0], linewidth=2, ls=':', color='grey', alpha=0.5)
# ax2.legend(['Predicted cases', 'Actual cases', 'Train-test split'], loc='upper left')
# ax2.set_xlabel("Day count ")
# ax2.set_ylabel("Logarithm new Cases")
#
# plt.suptitle(("newCases predictions based on Log-Lineal Regression for " + opted_country))
# The following is Lagged Linear prediction, the performance is not as quoted in the website, so there seems issues in this following code, reasons yet to be found out
train_days = int(dt_one_country.shape[0] * 0.95)
test_days = dt_one_country['Days Since'].iloc[-1] - train_days
lag_size = 30 #Lagged method as shown in the quoted website, keep lagged records (as columns) of 'lag_size' of new_cases
lagpred_data_features = dt_one_country.copy() #work with local copy, needed to do store inplace the predicted out and to compare with reference
lagpred_data_features = calculate_lag(lagpred_data_features, range(1, lag_size), 'new_cases') #update the new_cases_1,new_cases_2 etc columns
filter_col_new_cases = [col for col in lagpred_data_features if col.startswith('new_cases')] #use the additional lagging columns named as new_cases_1,new_cases_2, etc new_cases_29
lagpred_data_features[filter_col_new_cases] = lagpred_data_features[filter_col_new_cases].apply(
lambda x: np.log1p(x)) #Linear prediction with logarithm data
lagpred_data_features.replace([np.inf, -np.inf], 0, inplace=True)
lagpred_data_features.fillna(0, inplace=True)
start_fcst = 1 + lagpred_data_features['Days Since'].iloc[train_days] # prediction day 1
end_fcst = lagpred_data_features['Days Since'].iloc[-1] # last prediction day
for d in list(range(start_fcst, end_fcst + 1)): #do day by day fitting and prediction for each of the prediction days
X_train, Y_train_1, X_test = split_data_one_day(lagpred_data_features, d) #generate training and testing data for each day
model_1, pred_1 = lin_reg_lag(X_train, Y_train_1, X_test) #fit and predict for the day
lagpred_data_features.new_cases.iloc[d] = pred_1 #add the prediction data to the records
# Recompute lags
lagpred_data_features = calculate_lag(lagpred_data_features, range(1, lag_size), 'new_cases') #update the new_cases_1,new_cases_2 etc columns
lagpred_data_features.replace([np.inf, -np.inf], 0, inplace=True)
lagpred_data_features.fillna(0, inplace=True)
# print("Process for ", country_name, "finished in ", round(time.time() - ts, 2), " seconds")
predicted_data = lagpred_data_features.new_cases
real_data = dt_one_country.new_cases
# dates_list_num = list(range(0,len(dates_list)))
dates_list_num = dt_one_country['date']
# Plot results
model_scores.append(np.sqrt(mean_squared_error(real_data.iloc[train_days:], np.expm1(predicted_data.iloc[train_days:]))))
fig_LagPred = go.Figure()
fig_LagPred.add_trace(go.Scatter(x=dates_list_num, y=np.expm1(predicted_data),
mode='lines+markers', name="Prediction new Cases " + str(opted_country)))
fig_LagPred.add_trace(go.Scatter(x=dates_list_num, y=real_data,
mode='lines+markers', name="Validation Data for new Cases " + str(opted_country)))
fig_LagPred.add_vline(x=dates_list_num.iloc[start_fcst], line_dash="dash") # ,
# annotation=dict())#, annotation_position="top right")
# fig_LagPred.add_trace(go.Scatter(x=valid['date'], y=y_pred["Holt"],
# mode='lines+markers', name="Prediction of new Cases " + str(opted_country)))
fig_LagPred.update_layout(title="new Cases Linear Lagged Model Prediction",
xaxis_title="Date", yaxis_title="new Cases", legend=dict(x=0, y=1, traceorder="normal"))
# fig_LagPred.show()
# fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(15,6))
#
# ax1.plot(dates_list_num, np.expm1(predicted_data))
# ax1.plot(dates_list_num, real_data)
# ax1.axvline(dates_list_num.iloc[start_fcst], linewidth=2, ls = ':', color='grey', alpha=0.5)
# ax1.legend(['Predicted cases', 'Actual cases', 'Train-test split'], loc='upper left')
# ax1.set_xlabel("Day count ")
# ax1.set_ylabel("new Cases")
#
# ax2.plot(dates_list_num, predicted_data)
# ax2.plot(dates_list_num, np.log1p(real_data))
# ax2.axvline(dates_list_num.iloc[start_fcst], linewidth=2, ls = ':', color='grey', alpha=0.5)
# ax2.legend(['Predicted cases', 'Actual cases', 'Train-test split'], loc='upper left')
# ax2.set_xlabel("Day count ")
# ax2.set_ylabel("Log new Cases")
# plt.suptitle(("ConfirmedCases predictions based on Log-Lineal Regression for "+country_name))
model_names = ["Lasso Lars Regression", "Polynomial Regression","Holts Linear Prediction","Linear Regression Lagged Model"] #use this score to compare predictors
model_summary = pd.DataFrame(zip(model_names, model_scores),
columns=["Model Name", "Root Mean Squared Error"]).sort_values(
["Root Mean Squared Error"])
print(model_summary)
return fig_LarsReg, fig_PolyReg, fig_Holt, fig_LagPred
# LassoLars Regression import numpy as np from sklearn import datasets from sklearn.linear_model import LassoLars # load the iris datasets dataset = datasets.load_diabetes() # fit a LASSO using LARS model to the data model = LassoLars(alpha=0.1) model.fit(dataset.data, dataset.target) print(model) # make predictions expected = dataset.target predicted = model.predict(dataset.data) # summarize the fit of the model mse = np.mean((predicted-expected)**2) print(mse) print(model.score(dataset.data, dataset.target))
def ProcessData(df,vect1,vect2,builder):
descriptionmatrix = vect1.transform([str(x) for x in df['titledescription'].values])
locationmatrix = vect2.transform([str(x) for x in df['locationfull'].values])
# x = build_design_matrices([builder], df, return_type='dataframe', NA_action=NAAction(on_NA='drop', NA_types=[]))
y = df['SalaryNormalized'].values
#x_combo = np.hstack([np.asarray(x[0]),descriptionmatrix.toarray(),locationmatrix.toarray()])
x_combo = np.hstack([descriptionmatrix.toarray(),locationmatrix.toarray()])
return (np.asarray(y), sparse.coo_matrix(x_combo))
train = PreProcess(pd.read_csv('train.csv'))
(vect1,vect2,builder) = InitializeTransformers(train)
(y, x) = ProcessData(train, vect1, vect2,builder)
(y_test, x_test) = ProcessData(PreProcess(pd.read_csv('solution.csv')),vect1,vect2,builder)
lasso = Lasso()
lasso.fit(x,y)
y_pred = lasso.predict(x_test)
lassolars = LassoLars(alpha=2)
lassolars.fit(x.toarray(),y)
lars_pred = lassolars.predict(x_test)
print np.sqrt(mean_squared_error(y_test, y_pred))
print r2_score(y_test,y_pred)
print np.sqrt(mean_squared_error(y_test,lars_pred))
print r2_score(y_test,lars_pred)
