boston_sklearn.py

from sklearn.linear_model import LinearRegression
from scipy.stats.stats import pearsonr

# split into data and label arrays 
y = boston_pd['target']
X = boston_pd.drop(['target'], axis=1)

# create training (~80%) and test data sets
X_train = X[:400]
X_test = X[400:]
y_train = y[:400]
y_test = y[400:]

# train a classifier 
lr = LinearRegression()
model = lr.fit(X_train, y_train)

# make predictions
y_pred = model.predict(X_test)

# error metrics
r = pearsonr(y_pred, y_test)
mae = sum(abs(y_pred - y_test))/len(y_test
print("R-sqaured: " + str(r[0]**2))
print("MAE: " + str(mae))

mae = sum(abs(y_pred - y_test))/len(y_test
=> needs trailing parenthesis.

Too lazy to do a PR.