D491 Introduction to Analytics - Set 4 - Part 3

Correct Answer: They combine predictions from multiple models

Ensemble learning methods combine the predictions of multiple models to improve performance and reduce the risk of overfitting.

Correct Answer: To improve model performance by optimizing hyperparameters

Hyperparameter tuning involves adjusting the parameters that control the learning process (such as learning rate or number of layers) to improve model performance.

Correct Answer: Convolutional Neural Networks (CNNs)

CNNs are specialized for image recognition tasks because they capture spatial hierarchies in images through convolutional layers.

Correct Answer: To tune hyperparameters and prevent overfitting

The validation set is used during training to fine-tune hyperparameters and prevent the model from overfitting on the training data.

Correct Answer: Term Frequency-Inverse Document Frequency (TF-IDF)

TF-IDF is a feature extraction technique that reflects how important a word is in a document relative to other documents in a collection.

Correct Answer: The number of training examples processed before updating the model parameters

Batch size refers to the number of training examples used in one iteration to update the models weights during training.

Correct Answer: Random forest

Random forest is commonly used for fraud detection because of its ability to handle large datasets and classify data into various categories with high accuracy.

Correct Answer: Reinforcement learning

Reinforcement learning is used in autonomous driving systems where an agent learns to make decisions through interactions with the environment, receiving feedback in the form of rewards or penalties.

Correct Answer: To assess how well the model generalizes to new data

Cross-validation is used to evaluate how well a model generalizes by partitioning the data into multiple subsets and testing the model on different combinations of these subsets.

Correct Answer: To normalize the range of independent variables

Feature scaling ensures that all features have a similar scale, which can improve the performance of many machine learning algorithms, such as gradient descent.