Machine Learning Quiz

What is the purpose of sampling smaller training sets in large datasets?

• To improve model accuracy

• To speed up training time

• To penalize complex models

• To increase data variability

Why is it important to analyze the most significant variables for each algorithm?

• To reduce the training time

• To understand the model's decision-making process

• To avoid overfitting

• To increase model complexity

What is the role of feature selection and engineering in the modelling process?

• To add more features to the dataset

• To remove redundant or irrelevant features

• To simplify the algorithm

• To increase the dataset size

What is the primary purpose of the ColumnTransformer in the given practical?

• To scale numerical features

• To apply different preprocessing steps to numerical and categorical features

• To encode categorical features

• To handle missing values

Why might it be beneficial to use a pipeline in model training?

• To automate the preprocessing steps

• To reduce the amount of code

• To ensure reproducibility of results

• All of the above

What is the main goal of K-fold cross-validation?

• To reduce the amount of data needed

• • To increase the training speed

• To provide a more reliable estimate of model performance

• • To simplify the model

Which of the following preprocessing steps is typically applied to numerical data?

• One-Hot Encoding

• Simple Imputer

• Label Encoding

• Text Vectorization

What techniques are used to fine-tune hyperparameters of a model?

• Grid Search

• Random Search

• Manual Tuning

• Genetic Algorithms

Why is it important to split your data into training and test sets?

• To use all data for training

• To ensure the model performs well on unseen data

• To reduce computation time

• To simplify the model

What does the mean_squared_error function measure in regression models?

• The average error rate

• The average squared difference between predicted and actual values

• The average absolute difference between predicted and actual values

• The percentage of correctly predicted values