def gradient_boosting_model(_remove_extra_classes=False, metric="accuracy"):
print("###Running Gradient Boosting model| Limited dataset:{}".format(
_remove_extra_classes))
dt = bdt.getOneHotEncodedDataset(
remove_extra_classes=_remove_extra_classes)
X = dt.iloc[:, 1:31]
y = dt.iloc[:, 31]
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42,
stratify=y)
train = bdt.getBalancedDataset(X_train, y_train)
X_train = train.iloc[:, :-1]
y_train = train.iloc[:, -1]
gb_clf = GradientBoostingClassifier(n_estimators=100,
learning_rate=0.5,
max_depth=3,
random_state=42)
gb_clf.fit(X_train, y_train)
scores = cross_val_score(gb_clf, X_test, y_test, cv=5, scoring=metric)
y_pred = gb_clf.predict(X_test)
conf_matrix = confusion_matrix(y_test, y_pred)
tn, fp, fn, tp = conf_matrix.ravel()
print('{} testing : {:.3f} (+-{:.3f})'.format(metric, scores.mean(),
scores.std()))
print("confusion matrix:\n", conf_matrix)
print(
"True Negative:{0}, False Positive:{1} \nFalse Negative:{2}, True Positive:{3}"
.format(tn, fp, fn, tp))
print("###Finished running Gradient Boosting model")
return
def svm_model(_remove_extra_classes=False, metric="accuracy"):
print("###Running SVM model| Limited dataset:{}".format(
_remove_extra_classes))
dt = bdt.getOneHotEncodedDataset(
remove_extra_classes=_remove_extra_classes)
X = dt.iloc[:, 1:31]
y = dt.iloc[:, 31]
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42,
stratify=y)
train = bdt.getBalancedDataset(X_train, y_train)
X_train = train.iloc[:, :-1]
y_train = train.iloc[:, -1]
clf = svm.SVC(gamma='scale',
decision_function_shape='ovo',
C=1.0,
cache_size=200,
kernel='rbf')
clf.fit(X_train, y_train)
scores = cross_val_score(clf, X_test, y_test, cv=5, scoring=metric)
y_pred = clf.predict(X_test)
conf_matrix = confusion_matrix(y_test, y_pred)
tn, fp, fn, tp = confusion_matrix(y_test, y_pred).ravel()
print('{} testing : {:.3f} (+-{:.3f})'.format(metric, scores.mean(),
scores.std()))
print("confusion matrix:\n", conf_matrix)
print(
"True Negative:{0}, False Positive:{1} \nFalse Negative:{2}, True Positive:{3}"
.format(tn, fp, fn, tp))
print("###Finished running SVM model")
return
def mlp_model(_remove_extra_classes=False, metric="accuracy"):
print(
"Running MLP model| Limited dataset:{}".format(_remove_extra_classes))
dt = bdt.getOneHotEncodedDataset(
remove_extra_classes=_remove_extra_classes)
X = dt.iloc[:, 1:31]
y = dt.iloc[:, 31]
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42,
stratify=y)
train = bdt.getBalancedDataset(X_train, y_train)
X_train = train.iloc[:, :-1]
y_train = train.iloc[:, -1]
clf = MLPClassifier(solver='lbfgs',
alpha=1e-5,
hidden_layer_sizes=(90, 40),
random_state=1)
clf.fit(X_train, y_train)
scores = cross_val_score(clf, X_test, y_test, cv=5, scoring=metric)
y_pred = clf.predict(X_test)
conf_matrix = confusion_matrix(y_test, y_pred)
tn, fp, fn, tp = conf_matrix.ravel()
print('{} testing : {:.3f} (+-{:.3f})'.format(metric, scores.mean(),
scores.std()))
print("confusion matrix:\n", conf_matrix)
print(
"True Negative:{0}, False Positive:{1} \nFalse Negative:{2}, True Positive:{3}"
.format(tn, fp, fn, tp))
print("###Finished running MLP model")
return
def main():
dt = bdt.getOneHotEncodedDataset()
#dt = bdt.getBalancedDataset(dataset)
X = dt.iloc[ : , 1:31 ]
y = dt.iloc[ : , 31 ]
subsets = [
["IL-10 -592=CA","TNF-308=GG"],
["MBL -221=YX", "IL-10 -819=CT", "TNF-308=GG"],
["TNF-308=GG"],
["PTX3 rs2305619=GG", "MPO C-463T=GG"],
["PTX3 rs2305619=AA", "IL-10 -592=CA"],
["PTX3 rs2305619=AA"],
["IL-10 -819=CT", "MPO C-463T=GG"],
["PTX3 rs1840680=AA", "IL-28b rs12979860=CT"],
["MPO C-463T=GG"],
["PTX3 rs1840680=AA", "MBL -221=XX"]
]
print("###Running Experinmento1:")
for sset in subsets:
X_sset = X[sset]
#print(X_sset.columns)
X_train, X_test, y_train, y_test = train_test_split(X_sset, y, test_size=0.33, random_state=100, stratify=y)
gb_clf = GradientBoostingClassifier(n_estimators=100, learning_rate=0.5, max_depth=3, random_state=100)
gb_clf.fit(X_train, y_train)
scores = cross_val_score(gb_clf, X_train, y_train, cv=5, scoring='accuracy')
print("Score: {0:.3f} +-({1:.3f}))".format(scores.mean(), scores.std()))
y_pred = gb_clf.predict(X_test)
conf_matrix = confusion_matrix(y_test, y_pred)
tn, fp, fn, tp = conf_matrix.ravel()
print("confusion matrix:", conf_matrix)
print("True Negative:{0}, False Positive:{1} \nFalse Negative:{2}, True Positive:{3}".format(tn, fp, fn, tp))
print("###Finished Experinmento1")
return