terrah27

terrah27

# Separate input features and target
y = df.Class
X = df.drop('Class', axis=1)

# setting up testing and training sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=27)

# DummyClassifier to predict only target 0
dummy = DummyClassifier(strategy='most_frequent').fit(X_train, y_train)
dummy_pred = dummy.predict(X_test)

terrah27

# Modeling the data as is
# Train model
lr = LogisticRegression(solver='liblinear').fit(X_train, y_train)
 
# Predict on training set
lr_pred = lr.predict(X_test)

# Checking accuracy
accuracy_score(y_test, lr_pred)
    0.9992

terrah27

# f1 score
f1_score(y_test, lr_pred)
    0.7522
  
# recall score
recall_score(y_test, lr_pred)
    0.6439

terrah27

from sklearn.ensemble import RandomForestClassifier

# train model
rfc = RandomForestClassifier(n_estimators=10).fit(X_train, y_train)

# predict on test set
rfc_pred = rfc.predict(X_test)

accuracy_score(y_test, rfc_pred)
    0.9995

terrah27

from sklearn.utils import resample

# Separate input features and target
y = df.Class
X = df.drop('Class', axis=1)

# setting up testing and training sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=27)

# concatenate our training data back together

terrah27

# trying logistic regression again with the balanced dataset
y_train = upsampled.Class
X_train = upsampled.drop('Class', axis=1)

upsampled = LogisticRegression(solver='liblinear').fit(X_train, y_train)

upsampled_pred = upsampled.predict(X_test)

# Checking accuracy
accuracy_score(y_test, upsampled_pred)

terrah27

# still using our separated classes fraud and not_fraud from above

# downsample majority
not_fraud_downsampled = resample(not_fraud,
                                replace = False, # sample without replacement
                                n_samples = len(fraud), # match minority n
                                random_state = 27) # reproducible results

# combine minority and downsampled majority
downsampled = pd.concat([not_fraud_downsampled, fraud])

terrah27

# trying logistic regression again with the undersampled dataset

y_train = downsampled.Class
X_train = downsampled.drop('Class', axis=1)

undersampled = LogisticRegression(solver='liblinear').fit(X_train, y_train)

undersampled_pred = undersampled.predict(X_test)

# Checking accuracy

terrah27

from imblearn.over_sampling import SMOTE

# Separate input features and target
y = df.Class
X = df.drop('Class', axis=1)

# setting up testing and training sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=27)

sm = SMOTE(random_state=27, ratio=1.0)