Quiz on non-linear regression

Check how much you remember from previous sections by answering the questions below.

Which of the following best describes the difference between a linear and non-linear model in machine learning??

✓A linear model assumes a direct proportionality between the input and output variables, while a non-linear model does not.

✗Linear models are more accurate than non-linear models in all cases.

✗Non-linear models always involve more complex mathematical calculations than linear models.

✗A non-linear model cannot be used for regression problems.

What is the purpose of splitting a dataset into training and testing sets?

✓To avoid overfitting by using different data for model training and evaluation.

✗To ensure that the model sees all the data at least once.

✗To reduce the size of the dataset to make the model run faster.

✗To increase the accuracy of the predictions by using smaller datasets.

Which of the following steps is performed immediately after training a machine learning model?

✗Splitting the dataset.

✗Evaluating the model using the test set.

✗Fine-tuning the model.

✓Predicting on unseen data.

If a regression model has an R² score of 0.85, what does this indicate about the model’s performance?

✗The model has an error rate of 85%.

✗The model is overfitting the training data.

✗The model’s predictions are accurate by 85% on average.

✓The model explains 85% of the variance in the target variable.

Which of the following evaluation metrics penalizes large prediction errors more heavily?

✗Mean Absolute Error (MAE)

✗R-squared (R²)

✓Root Mean Squared Error (RMSE)

✗None of the above

What is the main benefit of using cross-validation over a simple train-test split?

✗Cross-validation ensures that the model will be 100% accurate.

✓Cross-validation helps to better estimate model performance by using multiple training-test splits.

✗Cross-validation makes the training process faster.

✗Cross-validation can only be used for linear models.

What is the main reason for computing spatially stratified metrics in model evaluation?

✗To get global evaluation scores for the entire dataset.

✓To account for regional differences and compute performance metrics for each geographical subset.

✗To increase the accuracy of the model’s predictions.

✗To ensure that the model is overfitting.

What does the term “spatial leakage” refer to in spatial cross-validation?

✗The leakage of information between temporal datasets.

✓When spatial proximity between training and test sets violates the assumption of independence.

✗The leakage of features from test data into the training data.

✗The violation of model performance evaluation due to overfitting.

In a Local Indicators of Spatial Association (LISA) analysis, what does a “High-High” cluster of residuals indicate?

✗A region with consistently high error rates.

✗A region with over-predicted values.

✓A region where the model consistently under-predicts.

✗A region with low error rates.

What does feature importance represent in tree-based models such as Gradient Boosting or Random Forest?

✗The statistical significance of each feature.

✗The accuracy of each feature’s predictions.

✗The raw coefficients assigned to each feature.

✓The relative contribution of each feature in reducing the model’s loss function.