def validation_curve_classifier(path="images/validation_curve_classifier.png"):
data = pd.read_csv(os.path.join(FIXTURES, "game", "game.csv"))
target = "outcome"
features = [col for col in data.columns if col != target]
X = pd.get_dummies(data[features])
y = data[target]
_, ax = plt.subplots()
cv = StratifiedKFold(12)
param_range = np.logspace(-6, -1, 12)
oz = ValidationCurve(
SVC(),
ax=ax,
param_name="gamma",
param_range=param_range,
logx=True,
cv=cv,
scoring="f1_weighted",
n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
def yellow_brick_validation_curve(model, x, y, cpu_count, cv_count, param,
scoring_metric):
"""
"""
from yellowbrick.model_selection import LearningCurve
from sklearn.model_selection import StratifiedKFold
# Create the learning curve visualizer
cv = StratifiedKFold(n_splits=cv_count)
# Validation Curve
mpl.rcParams['axes.prop_cycle'] = cycler('color', ['purple', 'darkblue'])
fig = plt.gcf()
fig.set_size_inches(10, 10)
ax = plt.subplot(411)
viz = ValidationCurve(model,
n_jobs=cpu_count,
ax=ax,
param_name=param,
param_range=np.arange(1, 11),
cv=cv,
scoring=scoring_metric)
# Fit and poof the visualizer
viz.fit(x, y)
viz.show()
def validation_curve_classifier_alt(
path="images/validation_curve_classifier_alt.png"):
data = pd.read_csv(os.path.join(FIXTURES, "game", "game.csv"))
target = "outcome"
features = [col for col in data.columns if col != target]
X = pd.get_dummies(data[features])
y = data[target]
_, ax = plt.subplots()
cv = StratifiedKFold(4)
param_range = np.arange(3, 20, 2)
oz = ValidationCurve(
KNeighborsClassifier(),
ax=ax,
param_name="n_neighbors",
param_range=param_range,
cv=cv,
scoring="f1_weighted",
n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
def plot_validation_curve(final_X, final_Y):
viz = ValidationCurve(DecisionTreeClassifier(),
param_name="max_depth",
param_range=np.arange(1, 30),
cv=10,
scoring="accuracy")
viz.fit(final_X, final_Y)
viz.poof()
def validation():
X, y = load_energy()
oz = ValidationCurve(
DecisionTreeRegressor(),
param_name="max_depth",
param_range=np.arange(1, 11),
cv=10,
scoring="r2",
ax=newfig(),
)
oz.fit(X, y)
savefig(oz, "validation_curve")
def validation_curve_sklearn_example(path="images/validation_curve_sklearn_example.png"):
digits = load_digits()
X, y = digits.data, digits.target
_, ax = plt.subplots()
param_range = np.logspace(-6, -1, 5)
oz = ValidationCurve(
SVC(), ax=ax, param_name="gamma", param_range=param_range,
logx=True, cv=10, scoring="accuracy", n_jobs=4
)
oz.fit(X, y)
oz.poof(outpath=path)
def draw_validation_curve(self,
param_name,
param_range,
cv,
logx=False,
scoring="accuracy",
n_jobs=5):
visualizer = ValidationCurve(self.model,
param_name=param_name,
param_range=param_range,
logx=logx,
cv=cv,
scoring=scoring,
n_jobs=n_jobs)
visualizer.fit(self.training_data, self.training_labels)
visualizer.poof()
def validation_curve_sklearn_example(
path="images/validation_curve_sklearn_example.png"):
digits = load_digits()
X, y = digits.data, digits.target
_, ax = plt.subplots()
param_range = np.logspace(-6, -1, 5)
oz = ValidationCurve(SVC(),
ax=ax,
param_name="gamma",
param_range=param_range,
logx=True,
cv=10,
scoring="accuracy",
n_jobs=4)
oz.fit(X, y)
oz.poof(outpath=path)
def generate_validation_curve(model, clf_name, param_name, param_range,
scoring, cv, dataset_name, X_train, y_train):
if 'svm' in clf_name or 'nn' == clf_name:
train_scores, test_scores = validation_curve(model,
X_train,
y_train,
param_name=param_name,
param_range=param_range,
scoring="accuracy",
n_jobs=8)
train_scores_mean = np.mean(train_scores, axis=1)
train_scores_std = np.std(train_scores, axis=1)
test_scores_mean = np.mean(test_scores, axis=1)
test_scores_std = np.std(test_scores, axis=1)
plt.title("Validation Curve with {}".format(clf_name))
plt.xlabel(param_name)
plt.ylabel("Score")
plt.semilogx(param_range,
train_scores_mean,
label="Training score",
marker='o',
color="#0272a2")
plt.semilogx(param_range,
test_scores_mean,
label="Cross-validation score",
marker='o',
color="#9fc377")
plt.legend(loc="best")
plt.savefig("results/{}_model_complexity_{}_{}.png".format(
clf_name, dataset_name, param_name))
plt.clf()
else:
viz = ValidationCurve(model,
param_name=param_name,
param_range=param_range,
scoring=scoring,
cv=cv)
viz.fit(X_train, y_train)
viz.show("results/{}_model_complexity_{}_{}.png".format(
clf_name, dataset_name, param_name))
plt.clf()
def validation_curve_classifier_alt(path="images/validation_curve_classifier_alt.png"):
data = pd.read_csv(os.path.join(FIXTURES, "game", "game.csv"))
target = "outcome"
features = [col for col in data.columns if col != target]
X = pd.get_dummies(data[features])
y = data[target]
_, ax = plt.subplots()
cv = StratifiedKFold(4)
param_range = np.arange(3, 20, 2)
oz = ValidationCurve(
KNeighborsClassifier(), ax=ax, param_name="n_neighbors",
param_range=param_range, cv=cv, scoring="f1_weighted", n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
def validation_curve_regressor(path="images/validation_curve_regressor.png"):
data = pd.read_csv(os.path.join(FIXTURES, "energy", "energy.csv"))
targets = ["heating load", "cooling load"]
features = [col for col in data.columns if col not in targets]
X = data[features]
y = data[targets[1]]
_, ax = plt.subplots()
param_range = np.arange(1, 11)
oz = ValidationCurve(
DecisionTreeRegressor(), ax=ax, param_name="max_depth",
param_range=param_range, cv=10, scoring="r2", n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
def validation_curve_classifier(path="images/validation_curve_classifier.png"):
data = pd.read_csv(os.path.join(FIXTURES, "game", "game.csv"))
target = "outcome"
features = [col for col in data.columns if col != target]
X = pd.get_dummies(data[features])
y = data[target]
_, ax = plt.subplots()
cv = StratifiedKFold(12)
param_range = np.logspace(-6, -1, 12)
oz = ValidationCurve(
SVC(), ax=ax, param_name="gamma", param_range=param_range,
logx=True, cv=cv, scoring="f1_weighted", n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
def validation_curve(model, X, y):
from yellowbrick.model_selection import ValidationCurve
from sklearn.model_selection import StratifiedKFold
# Create the validation curve visualizer
cv = StratifiedKFold(12)
# param_range = np.linspace(30.00, 300.00, num=50.00, dtype=np.float64)
param_range = np.logspace(30, 300, num=100, dtype=np.int32)
viz = ValidationCurve(
model,
param_name="n_estimators",
param_range=param_range,
logx=True,
cv=cv,
scoring="f1_weighted",
n_jobs=8,
)
viz.fit(X, y)
viz.poof()
def validation_curve(model, x, y, param, rang, cv):
"""
:param model: Modelo a ser avaliado.
:param x: Variáveis independentes de treino.
:param y: Variavel dependente de treino.
:param param: Parametro do modelo a ser avaliado.
:param rang: Espaço de hipotese do parametro que esta sendo avaliado.
:param cv: quantidade de splits para a cross validação.
:return: Viz das curvas de validação.
"""
viz = ValidationCurve(model,
param_name=param,
param_range=rang,
cv=cv,
scoring="roc_auc",
n_jobs=-1)
viz.fit(x, y)
viz.show()
def validation_curve_classifier_knn(path="images/validation_curve_classifier_knn.png"):
X, y = load_game()
X = OneHotEncoder().fit_transform(X)
_, ax = plt.subplots()
cv = StratifiedKFold(4)
param_range = np.arange(3, 20, 2)
print("warning: generating the KNN validation curve can take a very long time!")
oz = ValidationCurve(
KNeighborsClassifier(),
ax=ax,
param_name="n_neighbors",
param_range=param_range,
cv=cv,
scoring="f1_weighted",
n_jobs=8,
)
oz.fit(X, y)
oz.show(outpath=path)
def validation_curve_classifier_svc(path="images/validation_curve_classifier_svc.png"):
X, y = load_game()
X = OneHotEncoder().fit_transform(X)
_, ax = plt.subplots()
cv = StratifiedKFold(12)
param_range = np.logspace(-6, -1, 12)
print("warning: generating the SVC validation curve can take a very long time!")
oz = ValidationCurve(
SVC(),
ax=ax,
param_name="gamma",
param_range=param_range,
logx=True,
cv=cv,
scoring="f1_weighted",
n_jobs=8,
)
oz.fit(X, y)
oz.show(outpath=path)
def validation_curve_regressor(path="images/validation_curve_regressor.png"):
data = pd.read_csv(os.path.join(FIXTURES, "energy", "energy.csv"))
targets = ["heating load", "cooling load"]
features = [col for col in data.columns if col not in targets]
X = data[features]
y = data[targets[1]]
_, ax = plt.subplots()
param_range = np.arange(1, 11)
oz = ValidationCurve(
DecisionTreeRegressor(),
ax=ax,
param_name="max_depth",
param_range=param_range,
cv=10,
scoring="r2",
n_jobs=8,
)
oz.fit(X, y)
oz.poof(outpath=path)
# ### Validation Curve with C (metric: accuracy)
# In[59]:
# Plotting
from yellowbrick.model_selection import ValidationCurve
param_range = np.arange(0.001, 10)
viz = ValidationCurve(LogisticRegression(solver='liblinear'),
param_name='C',
param_range=param_range,
cv=5,
scoring="accuracy")
viz.fit(X_train, y_train)
viz.show()
# ### Validation Curve with C (metric: AUC score)
# In[60]:
from sklearn.metrics import roc_auc_score, roc_curve, confusion_matrix, classification_report, auc
C = [0.1, 0.5, 1, 1.5, 1.75, 2]
train_results = []
test_results = []
for c in C:
model = LogisticRegression(C=c, solver='liblinear', random_state=0)
model.fit(X_train, y_train)
train_pred = model.predict_proba(X_train)[:, 1]
tuned.fit(x_train, y_train)
return tuned.best_params_
# %%
dt = DTClassifier()
# %%
viz = ValidationCurve(
DecisionTreeClassifier(), param_name="max_depth",
param_range=np.arange(1, 21), cv=10, scoring="f1_weighted"
)
# Fit and show the visualizer
viz.fit(x_data,y_data)
viz.show()
# %%
viz = ValidationCurve(
DecisionTreeClassifier(), param_name="min_samples_leaf",
param_range=np.arange(1, 21), cv=10, scoring="f1_weighted"
)
# Fit and show the visualizer
viz.fit(x_data,y_data)
viz.show()
# %%
print("Balanced_accuracy:{:.4f}".format(svf))
#print("accuracy:{:.4f}".format(svf_acc))
data_score = pd.DataFrame(columns=['Commodity', 'score'])
data_score['Commodity'] = y_location_trains.columns
data_score['score'] = scores
print(data_score)
df_score.to_csv('/Users/monalisa/Downloads/mmai823-project-master/out/SVF_scores.csv')
# Vizualization Curve is better for SVM
from sklearn.model_selection import StratifiedKFold
from matplotlib import pyplot as plt
%matplotlib inline
plt.tight_layout()
cv = StratifiedKFold(12)
param_range = np.logspace(-6, -1, 12)
viz = ValidationCurve(
SVC(), param_name="gamma", param_range=param_range,
logx=True, cv=cv, scoring="roc_auc", n_jobs=8,
)
viz.fit(X_train, training_scores_encoded)
viz.show()
plt.title('Cross Validation Results')
plt.tight_layout()
plt.gcf().set_size_inches(10, 5)
plt.show()
# Train & Validation Curves mit yellowbricks
fig, ax = plt.subplots(figsize=(16, 9))
val_curve = ValidationCurve(
KNeighborsRegressor(),
param_name='n_neighbors',
param_range=n_neighbors,
cv=5,
scoring=rmse_score,
# n_jobs=-1,
ax=ax)
val_curve.fit(X, y)
val_curve.poof()
fig.tight_layout()
plt.show()
fig, ax = plt.subplots(figsize=(16, 9))
l_curve = LearningCurve(
KNeighborsRegressor(n_neighbors=best_k),
train_sizes=np.arange(.1, 1.01, .1),
scoring=rmse_score,
cv=5,
# n_jobs=-1,
ax=ax)
l_curve.fit(X, y)
l_curve.poof()
fig.tight_layout()
for ind, model in enumerate(models):
model.fit(x_train, y_train)
preds = model.predict(x_test)
for index, ax in enumerate(axes):
residuals_plot(model, x_test, preds, hist=False, ax=ax[index])
prediction_error(model, x_test, preds, ax=ax)
# Do some scoring on XGB estimators
# Validation curve
viz = ValidationCurve(XGBRegressor(objective="reg:squarederror"),
param_name="max_depth",
param_range=np.arange(1, 11),
cv=5,
scoring="r2")
viz.fit(x_train, y_train)
viz.show()
# Learning curve
model = XGBRegressor(objective="reg:squarederror")
viz_2 = LearningCurve(model, scoring="r2")
viz_2.fit(x_train, y_train)
viz_2.show()
model = RFECV(LassoCV(), cv=5, scoring='r2')
model.fit(x_train, y_train)
model.show()
"""
Section: 5
Time-Series Algorithms
"""
case_name = "mg_sizing_dataset_with_loc"
df = pd.read_csv("results/" + case_name + ".csv", sep=";|,", engine="python", index_col='index')
#df = df.loc[df['off-grid'] == 1]
X = df[features]
scaler.fit(X)
X = scaler.transform(X)
# X = pd.DataFrame(scaler.transform(X), index=X.index, columns=X.columns)
targets = ["PV","BAT","RBAT","INV","GEN","NPV"]
y = df[targets]
cv = StratifiedKFold(12)
param_range = np.arange(1, 30, 1)
cv = KFold(n_splits=12, random_state=40, shuffle=True)
viz = ValidationCurve(
KNeighborsRegressor(), param_name="n_neighbors", param_range=param_range, scoring="r2", cv=cv, n_jobs=8
)
viz.fit(X, y)
viz.show()
visualizer = LearningCurve(KNeighborsRegressor(), scoring='r2', random_state=2, cv=cv, shuffle=True)
visualizer.fit(X, y)
visualizer.show()
vis = CVScores(KNeighborsRegressor(), cv=cv, scoring='r2')
vis.fit(X, y) # Fit the data to the visualizer
vis.show()