Machine Learning Foundations

Question 1

Define:

feature engineering

Answer 1

The process of selecting, modifying, or creating input variables (features) that help improve the performance of a machine learning model.

Good feature engineering can significantly enhance the accuracy of a model.

Question 2

True or False:

Feature engineering is not necessary for machine learning models.

Answer 2

False

Effective feature engineering can greatly affect a model’s success by providing better inputs for learning.

Question 3

Fill in the blank:

The dataset is split into ______ and testing sets to evaluate a model’s performance.

Answer 3

training

Question 4

What is the purpose of the training dataset?

Answer 4

To teach a machine learning model by providing examples that the model can learn from.

Question 5

What is the purpose of the testing dataset?

Answer 5

To evaluate the performance of a trained machine learning model on new, unseen data.

Question 6

What is model evaluation in machine learning?

Answer 6

The process of assessing how well a machine learning model performs on a given dataset using various metrics.

Question 7

True or False:

A model that performs well on training data will always perform well on new data.

Answer 7

False

This is not always true due to overfitting, where the model learns the training data too well and fails to generalize to new data.

Question 8

What does overfitting mean?

Answer 8

A situation where a model learns the training data too closely, capturing noise as if it were a true pattern, which negatively impacts its performance on new data.

Question 9

What does underfitting mean?

Answer 9

A situation where a model is too simple to capture the underlying patterns in the data, resulting in poor performance on both training and testing data.

Question 10

Fill in the blank:

A common metric for evaluating classification models is ______.

Answer 10

accuracy

Question 11

What is accuracy in the context of model evaluation?

Answer 11

The proportion of correct predictions made by a machine learning model out of all predictions made.

Question 12

What is a limitation of using accuracy as an evaluation metric?

Answer 12

Accuracy can be misleading in imbalanced datasets where one class is much more common than the others.

Question 13

What is precision in model evaluation?

Answer 13

The ratio of true positive predictions to the total number of positive predictions made, measuring the accuracy of positive predictions.

Question 14

What is recall in model evaluation?

Answer 14

The ratio of true positive predictions to the total number of actual positive cases, measuring the model’s ability to find all the relevant cases.

Question 15

What do the F1 score represent?

Answer 15

A metric that combines precision and recall into a single score by taking their harmonic mean, used to balance both metrics when they are important.

Question 16

Which metric would you prioritize for a medical diagnosis model: precision or recall?

Answer 16

recall

Question 17

Why would you prioritize recall for a medical diagnosis model?

Answer 17

In medical diagnosis, it’s crucial to identify as many positive cases as possible to ensure patients receive necessary treatment, even if some false positives occur.

Question 18

What is cross-validation in the context of model evaluation?

Answer 18

A technique for assessing how a model will generalize to an independent dataset by training and testing the model multiple times with different data splits.

Question 19

Fill in the blanks:

______ ______ is the process of selecting the best features to use in a machine learning model.

Answer 19

Feature selection

Question 20

Why is feature selection important?

Answer 20

It helps reduce the complexity of a model, improves performance, and reduces the risk of overfitting by eliminating irrelevant or redundant features.

Question 21

How does feature scaling affect machine learning models?

Answer 21

Feature scaling standardizes the range of independent variables, improving the convergence speed of algorithms and ensuring features contribute equally to the results.

Question 22

What is the difference between normalization and standardization in feature scaling?

Answer 22

• Normalization scales the data to a range [0, 1].
• Standardization scales data to have a mean of 0 and a standard deviation of 1.

Question 23

What is a confusion matrix?

Answer 23

A table used to evaluate the performance of a classification model, showing the true versus predicted classifications.

Question 24

In a scenario where you need to recommend movies to users, which machine learning subfield would be most applicable?

Answer 24

Collaborative filtering, a technique within machine learning used in recommendation systems.

Question 25

Why is it important to have **separate training** and **testing datasets**?

Answer 25

To ensure that the performance evaluation of the model is **unbiased** and reflects its ability to **generalize to unseen data**.