What is a binary classification problem?
- A problem with two algorithms
- A problem with two classes or outcomes
- A problem with two input variables
- A problem with two solutions
A binary classification problem involves predicting one of two possible classes or outcomes. It's a specific type of classification where the target variable has only two possible values.
What are the potential issues with using a large value of K in KNN, and how can they be mitigated?
- All of the above
- Complexity, can be mitigated by feature selection
- Overfitting, can be mitigated by reducing dimensions
- Underfitting, can be mitigated by reducing K
A large value of K in KNN may cause underfitting. It can be mitigated by reducing K or using techniques like cross-validation.
In a situation where the MAE is significantly lower than the RMSE, what might this tell you about the distribution of the errors in your model?
- Errors are normally distributed; no impact on model
- Errors are uniformly distributed; no large outliers
- Many large errors, few small outliers
- Many small errors, few large outliers
When the Mean Absolute Error (MAE) is significantly lower than the Root Mean Squared Error (RMSE), it likely indicates that the model has many small errors and a few large outliers. RMSE, being sensitive to larger errors, would be higher, while MAE would be less impacted by those larger errors. An analysis of the residuals can further elucidate the nature of these errors.
In the context of SVM, the support vectors are the data points that lie closest to the _________.
- Data center
- Hyperplane
- Kernel
- Margin
Support vectors are the data points that lie closest to the hyperplane and influence its position and orientation.
Can you explain what separates support vectors from other data points in SVM?
- Cluster membership
- Color
- Distance to hyperplane
- Size
Support vectors are the data points closest to the hyperplane, and they determine its position in SVM.
How do interpretability and explainability vary between AI, Machine Learning, and Deep Learning?
- AI and Deep Learning are equally interpretable, Machine Learning is least
- AI is least interpretable, Machine Learning and Deep Learning are equally interpretable
- AI is most interpretable, Machine Learning is moderate, Deep Learning is least
- Machine Learning is most interpretable, AI is moderate, Deep Learning is least
Generally, AI techniques can vary in interpretability, traditional Machine Learning models tend to be more interpretable, and Deep Learning models are often the least interpretable due to their complexity.
How does the choice of loss function affect the learning process in a Machine Learning model?
- It defines the optimization algorithm
- It determines the learning rate
- It measures how well the model's predictions match the true values
- It selects the type of regularization
The loss function measures the discrepancy between the predicted values and the actual values, guiding the optimization process. Different loss functions can emphasize different aspects of the error, influencing how the model learns.
How do the hyperparameters in Ridge and Lasso affect the bias-variance tradeoff?
- Increase bias, reduce variance
- Increase both bias and variance
- No effect
- Reduce bias, increase variance
The hyperparameters in Ridge and Lasso control the regularization strength. Increasing them increases bias but reduces variance, helping to prevent overfitting.
How do multi-class classification problems differ from binary classification problems?
- Multi-class has more features
- Multi-class has multiple classes; binary has two
- Multi-class has two classes; binary has multiple
- No difference
Binary classification deals with two classes, while multi-class classification deals with more than two. Multi-class problems can be more complex and require different handling or algorithms compared to binary classification.
How would you handle a multi-class classification problem with LDA, and what is the importance of maximizing between-class variance in this context?
- Apply LDA separately to each class; no importance of between-class variance
- Apply LDA to all classes; maximize between-class variance for class separability
- Ignore between-class variance and focus on within-class variance
- Use another method entirely
For a multi-class classification problem, you would "apply LDA to all classes" and maximize between-class variance. This is essential for separating the classes from each other, which improves classification performance.