ML Classification Basics

What is the main goal of supervised learning in machine learning?

• To predict future outcomes without labeled data

• To analyze unstructured data without any guidance

• To learn a mapping from input variables to output variables based on labeled training data.

• To generate random patterns without any input

Answer explanation

To learn a mapping from input variables to output variables based on labeled training data.

What is a popular classification algorithm used in machine learning?

• Decision Tree

• Logistic Regression

• K-Nearest Neighbors

• Support Vector Machine (SVM)

Answer explanation

Decision Tree, K-Nearest Neighbors, and Support Vector Machine (SVM) are popular classification algorithms used in machine learning.

Explain the concept of overfitting in the context of student performance prediction.

• Overfitting is when a model learns the training data too slowly, causing poor performance.

• Overfitting is when a model learns the training data too well, including the noise and outliers, leading to poor generalization to new data.

• Overfitting is when a model learns the training data perfectly, leading to flawless generalization.

• Overfitting is when a model learns the training data too quickly, resulting in accurate predictions.

Answer explanation

Overfitting is when a model learns the training data too well, including the noise and outliers, leading to poor generalization to new data.

What are some common techniques used for feature selection in machine learning?

• Principal component analysis

• Stepwise regression

• Filter methods, Wrapper methods, Embedded methods

• Backward elimination

Answer explanation

The common techniques used for feature selection in machine learning include Filter methods, Wrapper methods, and Embedded methods.

How does underfitting affect the performance of a machine learning model?

• Underfitting has no impact on the model's performance

• Underfitting results in overfitting of the model

• Underfitting improves the performance of the model by simplifying it

• Underfitting leads to poor performance of the machine learning model due to its inability to capture the complexity of the data.

Answer explanation

Underfitting leads to poor performance of the machine learning model due to its inability to capture the complexity of the data.

What is the difference between precision and recall in medical diagnosis?

• Precision is more important than recall in medical diagnosis tasks.

• Precision measures the ability to find all positive cases, while recall measures the accuracy of positive predictions.

• Precision focuses on the accuracy of positive predictions, while recall focuses on finding all positive cases.

• Precision is the ratio of true positive predictions to all positive predictions, while recall is the ratio of true positive predictions to all actual positive cases.

Answer explanation

Precision focuses on the accuracy of positive predictions, while recall focuses on finding all positive cases.

Give an example of a binary classification problem in the context of medical diagnosis.

• Predicting whether a patient has diabetes or not

• Predicting whether a tumor is malignant or benign

• Determining the blood type of an individual

• Identifying the breed of a dog

Answer explanation

The correct choice is predicting whether a tumor is malignant or benign, as it involves classifying tumors into two categories based on their nature, making it a binary classification problem in medical diagnosis.

How can regularization help prevent overfitting in machine learning models?

• Regularization adds a penalty term to the loss function, discouraging complex models and promoting generalization.

• Regularization has no impact on overfitting

• Regularization increases model complexity to fit the training data better

• Regularization reduces the number of training epochs

Answer explanation

Regularization adds a penalty term to the loss function, discouraging complex models and promoting generalization.

What is the importance of cross-validation in machine learning?

• Cross-validation is a technique to reduce the number of features in a dataset

• Cross-validation is only applicable to linear regression models

• Cross-validation is used to assess model performance, detect overfitting, and select the best model.

• Cross-validation is used to generate new data for training models

Answer explanation

Cross-validation is used to assess model performance, detect overfitting, and select the best model.

Discuss the bias-variance tradeoff in the context of loan approval models.

• Bias and variance are not related to loan approval model performance.

• The bias-variance tradeoff is about managing the tradeoff between underfitting (high bias) and overfitting (high variance) in loan approval models.

• Increasing bias always leads to better loan approval model performance.

• The bias-variance tradeoff only affects regression models, not loan approval models.

Answer explanation

The bias-variance tradeoff is about managing the tradeoff between underfitting (high bias) and overfitting (high variance) in loan approval models.