Linear Regression, Logistic Regression, PCA Quiz

What is the primary purpose of the learning rate α (alpha) in gradient descent algorithms?

• To set the initial weights of the model

• To control the step size during the weight update process

• To determine the stopping criteria for the algorithm

• • To calculate the cost function at each iteration

In the context of gradient descent, what does "convergence" typically refer to?

• The process of overfitting to the training data

• The flattening of the learning curve over time

• The point at which the cost function reaches a minimum value

• The phase where the model generalizes well to unseen data

Which of the following best describes the primary objective of linear regression?

• Classifying data into discrete categories

• Predicting a continuous output variable based on one or more input features

• Finding the mode of a dataset

• Clustering similar data points together

Why can't the Least Squared Sum cost function be used for Logistic Regression optimization?

• It overfits the model

• It can result in multiple local minimums

• It is computationally expensive

• It is non-differentiable

Which cost function is commonly used for Logistic Regression?

• Mean Squared Error

• Hinge Loss

• Log Loss

• Cross-Entropy Loss

What is a common threshold value for classification in Logistic Regression?

• 0.25

• 0.5

• 0.75

• 1.0

What is the primary goal of Principal Component Analysis (PCA)?

• Data Augmentation

• Data Transformation

• Dimensionality Reduction

• Data Smoothing

What does the covariance matrix represent in the context of PCA?

• The mean of each feature

• The linear relationship between each pair of features

• The variance of each feature

• The skewness of each feature

Which of the following steps is NOT part of PCA using the eigenvectors and eigenvalues method?

• Standardization

• Computation of Covariance matrix

• Singular Value Decomposition

• Compute eigenvectors and eigenvalues

Why is it crucial to standardize the dataset before conducting Principal Component Analysis (PCA)?

• To ensure features have comparable scales

• To transform the data distribution to Gaussian

• To sparsify the data for easier computation

• To artificially increase the dimensionality of the dataset