Tuning contamination

Finally, it is time to tune the notorious contamination parameter. The evaluate_outlier_classifier and evaluate_regressor functions from the video are already loaded for you. You can inspect them below.

def evaluate_outlier_classifier(model, data):
    # Get labels
    labels = model.fit_predict(data)

    # Return inliers
    return data[labels == 0]

def evaluate_regressor(inliers):
    X = inliers.drop("price", axis=1)
    y = inliers[['price']]
    X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=10)

    lr = LinearRegression()
    lr.fit(X_train, y_train)

    preds = lr.predict(X_test)
    rmse = mean_squared_error(y_test, preds, squared=False)

    return round(rmse, 3)

You will be using a sample of the US Airbnb Listings dataset, which has already been loaded as airbnb_df.

Create a list called contaminations that contains four values, 0.07, 0.1, 0.15, 0.25, and create an empty dictionary called scores to store the RMSE scores.

Exercise

Tuning contamination

Instructions 1/4