BAFri4LecTopic1012

What is the primary purpose of ensemble methods in machine learning?

• Reducing the number of features

• Combining multiple models to improve predictions

• Reducing computational complexity

• Increasing the size of training datasets

Which of the following is a boosting algorithm?

• Random Forest

• Gradient Boosting

• K-means Clustering

• Decision Trees

Random Forest can be described as:

• A collection of independent decision trees

• A neural network model

• A preprocessing tool for feature selection

• A linear regression model

Which of the following metrics is most important for fraud detection?

• R-squared

• Precision and Recall

• Mean Absolute Error

• Explained Variance

What is the primary advantage of deep learning over traditional machine learning?

• Ability to handle large, unstructured datasets

• Faster training times

• Lower computational cost

• Independence from hardware requirements

In a neural network, what is the purpose of the activation function?

• To standardize input data

• To introduce non-linearity into the model

• To compute the error gradient

• To reduce model overfitting

Which neural network layer type is primarily used for image data?

• Fully connected layer

• Convolutional layer

• Recurrent layer

• Pooling layer

Backpropagation in a neural network is used to:

• Predict test data labels

• Update weights based on loss gradients

• Initialize model weights

• Increase model accuracy through data augmentation

TensorFlow and Keras are commonly used for:

• Statistical data analysis

• Time series forecasting

• Deep learning model development

• Data cleaning

A hyperparameter in machine learning is:

• A parameter learned by the model during training

• A user-specified setting to configure the model

• Always fixed and cannot be tuned

• The output of the model