def OP_train(self):
OP_var = [1, 4, 5, 6, 7, 8, 9, 10, 11]
O_tar = self.Var[14]
P_tar = self.Var[15]
OP_train = self.data[self.Var[4:7]].values
OP_key = [i for i in OP_var if i in self.category_key]
for k in OP_key:
OP_train = np.concatenate(
(OP_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
O_label = np.argmax(
make_dummy(self.data[O_tar], self.dummy_dict[O_tar]), 1)
P_label = np.argmax(
make_dummy(self.data[P_tar], self.dummy_dict[P_tar]), 1)
O_label = make_dummy(self.data[O_tar], self.dummy_dict[O_tar])
P_label = make_dummy(self.data[P_tar], self.dummy_dict[P_tar])
O_rf = RandomForestClassifier(n_estimators=100)
O_rf.fit(X=OP_train, y=O_label)
P_rf = RandomForestClassifier(n_estimators=100)
P_rf.fit(X=OP_train, y=P_label)
with open("./Model/O_model.pickle", "wb") as f:
pickle.dump(O_rf, f)
with open("./Model/P_model.pickle", "wb") as f:
pickle.dump(P_rf, f)
def B_train(self):
B_var = [0, 2, 3, 4, 5, 6, 7, 8, 13, 14, 15]
B_tar = self.Var[1]
B_train = self.data[self.Var[2:7]].values
B_key = [i for i in B_var if i in self.category_key]
for k in B_key:
B_train = np.concatenate(
(B_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
B_label = np.argmax(
make_dummy(self.data[B_tar], self.dummy_dict[B_tar]), 1)
B_xgb = xgb.XGBClassifier()
B_xgb.fit(X=B_train, y=B_label)
with open("./Model/B_model.pickle", "wb") as f:
pickle.dump(B_xgb, f)
def A_train(self):
A_var = [1, 5, 6, 7, 8, 9, 10, 11, 12, 15]
A_tar = self.Var[0]
A_train = self.data[self.Var[5:7]].values
A_key = [i for i in A_var if i in self.category_key]
for k in A_key:
A_train = np.concatenate(
(A_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
A_label = np.argmax(
make_dummy(self.data[A_tar], self.dummy_dict[A_tar]), 1)
A_xgb = xgb.XGBClassifier()
A_xgb.fit(X=A_train, y=A_label)
with open("./Model/A_model.pickle", "wb") as f:
pickle.dump(A_xgb, f)
def K_train(self):
K_var = [0, 2, 3, 4, 5, 6, 7, 8, 13, 14, 15]
K_tar = self.Var[10]
K_train = self.data[self.Var[2:7]].values
K_key = [i for i in K_var if i in self.category_key]
for k in K_key:
K_train = np.concatenate(
(K_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
K_label = np.argmax(
make_dummy(self.data[K_tar], self.dummy_dict[K_tar]), 1)
K_xgb = xgb.XGBClassifier()
K_xgb.fit(X=K_train, y=K_label)
with open("./Model/K_model.pickle", "wb") as f:
pickle.dump(K_xgb, f)
def E_train(self):
E_var = [1, 7, 9, 10, 11, 12, 14, 15]
E_tar = self.Var[4]
E_train = self.data[self.Var[2:4] + self.Var[5:7]].values
E_key = [i for i in E_var if i in self.category_key]
for k in E_key:
E_train = np.concatenate(
(E_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
E_xgb = xgb.XGBClassifier()
E_xgb.fit(X=E_train, y=self.data[E_tar])
with open("./Model/E_model.pickle", "wb") as f:
pickle.dump(E_xgb, f)
def G_train(self):
G_var = [0, 1, 2, 5, 7, 9, 10, 11, 13, 14, 15]
G_tar = self.Var[6]
G_train = self.data[[self.Var[i] for i in [2, 5]]].values
G_key = [i for i in G_var if i in self.category_key]
for k in G_key:
G_train = np.concatenate(
(G_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
G_rf = RandomForestClassifier()
G_rf.fit(G_train, self.data[G_tar])
with open("./Model/G_model.pickle", "wb") as f:
pickle.dump(G_rf, f)
def C_train(self):
C_var = [9, 10, 11, 12]
C_tar = self.Var[2]
df = self.data[[self.Var[i] for i in C_var]]
c_train = self.data[self.Var[14:15]].values
for k in C_var:
c_train = np.concatenate(
(c_train,
make_dummy(self.data[self.Var[k]],
self.dummy_dict[self.Var[k]])),
axis=1)
tmp = pd.DataFrame(c_train)
del tmp[0]
clf = RandomForestClassifier(n_estimators=100)
clf.fit(tmp, self.data[C_tar])
with open("./Model/C_model.pickle", "wb") as f:
pickle.dump(clf, f)