def rfecv_credit_example(image="rfecv_credit.png"):
X, y = load_credit()
_, ax = plt.subplots()
cv = StratifiedKFold(5)
oz = RFECV(RandomForestClassifier(), ax=ax, cv=cv, scoring="f1_weighted")
oz.fit(X, y)
oz.poof(outpath=os.path.join(IMAGES, image))
def rfecv_sklearn_example(image="rfecv_sklearn_example.png"):
X, y = make_classification(
n_samples=1000,
n_features=25,
n_informative=3,
n_redundant=2,
n_repeated=0,
n_classes=8,
n_clusters_per_class=1,
random_state=0,
)
_, ax = plt.subplots()
oz = RFECV(SVC(kernel="linear", C=1), ax=ax)
oz.fit(X, y)
oz.poof(outpath=os.path.join(IMAGES, image))
# 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
"""
# Fitting ARIMA
# Original Series
# plt.rcParams.update({'figure.figsize':(9,7), 'figure.dpi':120})
fig, axes = plt.subplots(3, 1, sharex=True)
plot_acf(main_data.traffic_volume, ax=axes[0])
# 1st Differencing
plot_acf(main_data.traffic_volume.diff(), ax=axes[1])
def rfecv():
X, y = load_credit()
model = RandomForestClassifier(n_estimators=10)
oz = RFECV(model, cv=3, scoring="f1_weighted", ax=newfig())
oz.fit(X, y)
savefig(oz, "rfecv_sklearn_example")
y = fifa_semNan.Overall # In[24]: #instancia a classe reg = LinearRegression() # In[25]: #Ranking das features com RFE com cross validation #O ponto tracejado representa o score maximo com 27 features rfecv = RFECV(reg,step=1, cv=3) rfecv.fit(X,y) rfecv.show() # In[26]: #lista de features utilizadas pelo modelo list(zip(X.columns, rfecv.support_)) # In[27]: #seleciona 5 features para o modelo utilizando somente RFE
"auc_roc_ADAboost": 5,
"auc_roc_perceptron": 6,
"auc_roc_QDA": 7,
"auc_roc_linearSVC": 8,
"auc_roc_dtree": 9,
"auc_roc_bernoulliNB": 10,
"auc_roc_LDA": 11,
"auc_roc_gaussianNB": 12
}
df_results = df_results.replace({"Best model ROC": model_dict})
y = df_results["Best model ROC"]
cv = StratifiedKFold(2)
visualizer = RFECV(RandomForestClassifier(n_estimators=10),
cv=cv,
scoring='accuracy')
visualizer.fit(X, y) # Fit the data to the visualizer
visualizer.show()
print("Optimal number of features : %d" % visualizer.n_features_)
print(visualizer.ranking_)
print(visualizer.estimator_.feature_importances_)
index_list = []
for index, value in enumerate(visualizer.ranking_):
if value == 1:
index_list.append(index)
else:
pass
max_leaf_nodes=None,
min_impurity_decrease=0,
min_impurity_split=None,
min_samples_leaf=1,
min_samples_split=2,
min_weight_fraction_leaf=0,
n_jobs=None,
oob_score=False,
random_state=random_state,
max_depth=7,
max_features="log2",
verbose=0,
warm_start=False)
rfecv = RFECV(estimator=classifier,
step=1,
cv=StratifiedKFold(3),
scoring='f1_macro')
print(classifier)
print(rfecv)
# endregion
# region Train Model and plot importance
rfecv.fit(X_train, Y_train.ravel())
rfecv.show()
# endregion
visualizer.show()
# 学习率
visualizer = LearningCurve(model, scoring='f1_weighted')
visualizer.fit(X_train, y_train)
visualizer.show()
# 交叉验证
visualizer = CVScores(model, cv=5, scoring='f1_weighted')
visualizer.fit(X_train, y_train)
visualizer.show()
# 特征重要性
visualizer = FeatureImportances(model)
visualizer.fit(X_train, y_train)
visualizer.show()
# 特征递归消减
visualizer = RFECV(model, cv=5, scoring='f1_weighted')
visualizer.fit(X_train, y_train)
visualizer.show()
# 特征选择
visualizer = ValidationCurve(model,
param_name="max_depth",
param_range=np.arange(1, 11),
cv=5,
scoring="f1_weighted")
visualizer.fit(X_train, y_train)
visualizer.show()
.format(np.mean(precisions), np.std(precisions)))
CVlog(X, 'X')
#X:
#Avec threshold = 0.5 La moyenne des recalls pendant la cross validation est: 0.589 (~0.053)
#Avec threshold = 0.5 La moyenne des precisions pendant la cross validation est: 0.151 (~0.024)
#CVlog(X,'X',0.81)
#Avec threshold = 0.81 La moyenne des recalls pendant la cross validation est: 0.500 (~0.040)
#Avec threshold = 0.81 La moyenne des precisions pendant la cross validation est: 0.448 (~0.034)
##########################################################################
############# Réduction de features #############
# Cette étape peut prendre plusieurs heures (en fonction de l'ordinateur)
lr = LogisticRegression()
rfecv = RFECV(lr, step=1, cv=5, scoring='accuracy')
selector = rfecv.fit(X, Y)
print("The best descriptors after RFECV({}) are: \n {}".format(
rfecv.n_features_, X.columns[[
index for index, value in enumerate(list(rfecv.support_))
if value == True
]]))
X_bestFeatures = X.loc[:, X.columns[[
index for index, value in enumerate(list(rfecv.support_)) if value == True
]]]
CVlog(X_bestFeatures, 'X_bestFeatures')
#Avec threshold = 0.5 La moyenne des recalls pendant la cross validation est: 0.594 (~0.042)
#Avec threshold = 0.5 La moyenne des precisions pendant la cross validation est: 0.149 (~0.026)
##########################################################################
# def check_model(c, n, X, y, X_test, y_test, class_names, outdir):
# model = svm.LinearSVC(class_weight='balanced', dual=False, max_iter=10000, C=c)
# rfe = RFE(model, n_features_to_select=n)
# fit = rfe.fit(X,y)
# y_predict = fit.predict(X_test)
# predict_df = pd.DataFrame(y_predict.tolist())
# predict_df.to_csv(outdir + '/predict_label.csv', sep='\t', index=False)
# test = [[1e-3,12],[1e-6,7],[1.0,32],[1e-6,27]]
# for c,n in test:
# print(str(c)+ '-' +str(n))
# this_out = 'figures/eva/c' + str(c) + '_n' + str(n)
# check_model(c=c, n=n, outdir=this_out, X=X, y=y, X_test=X_test, y_test=y_test,
# class_names=class_names)
## plot RFECV for LinearSVC
for c in [1e-6, 1e-3, 1]:
model = svm.LinearSVC(class_weight='balanced', dual=False, max_iter=10000, C=c)
viz = RFECV(model, scoring='f1_weighted')
viz.fit(X, y)
viz.show(outpath='figures/linear_svc_rfecv.pdf')