Classification: sorting into groups quiz

Question 1

What is classification in machine learning?

Accepted Answer

Assigning data to predefined categories or classes

Answer

Predicting a number on a continuous scale, like tomorrow's temperature

Answer

Grouping data into clusters without using any labels

Answer

Deleting duplicate records from a dataset

Question 2

What is a 'label' or 'class' in a classification problem?

Accepted Answer

The correct category assigned to each training example

Answer

A column that has been removed from the dataset

Answer

A Python variable that stores numbers

Answer

A type of input feature used by the model

Question 3

What is supervised learning, and how does classification fit into it?

Accepted Answer

Learning from examples that already have correct labels

Answer

Learning without any data at all

Answer

Learning by trial and error using rewards and penalties

Answer

Learning by grouping similar items together without labels

Question 4

What is a 'feature' in a classification problem, and why does it matter?

Accepted Answer

A measurable property of the data used to make classification decisions

Answer

The final label predicted by the model

Answer

A type of mistake the model makes during training

Answer

The overall test accuracy of the trained model

Question 5

What does 'binary classification' mean?

Accepted Answer

Sorting every data point into exactly two possible classes

Answer

Classifying data into more than ten categories

Answer

Using binary numbers (0s and 1s) to store the dataset

Answer

Running the classification algorithm twice to improve accuracy

Question 6

What is a training dataset in the context of classification?

Accepted Answer

The collection of labeled examples a classifier learns from

Answer

A dataset set aside only for final performance reporting

Answer

A dataset that contains no labels at all

Answer

Data gathered after the model has been deployed to users

Question 7

Why do we keep a test dataset completely separate from the training data?

Accepted Answer

To check performance on unseen data and avoid inflated scores

Answer

To give the model extra examples to learn from

Answer

To store examples that were too noisy to use in training

Answer

To reduce the total size of the training set

Question 8

What is a decision tree classifier, and how does it make a prediction?

Accepted Answer

An algorithm that asks a series of yes/no questions about features to reach a class prediction

Answer

A chart that shows the neural connections in a network

Answer

A sorted list of all possible class labels

Answer

A method that groups similar data points into clusters without labels

Question 9

What is accuracy in classification, and when can it be misleading?

Accepted Answer

The fraction of correct predictions out of all predictions made

Answer

The speed at which the model makes predictions on new data

Answer

The number of features used by the model

Answer

The size of the training dataset used to train the model

Question 10

How does the K-Nearest Neighbors (KNN) algorithm classify a new data point?

Accepted Answer

It classifies a point based on the majority label of its K nearest training examples

Answer

It builds a tree of yes/no questions to classify data

Answer

It uses random weights to sort data into groups

Answer

It removes outliers from the dataset before making a prediction

Question 11

What is overfitting in a classification model, and how can you detect it?

Accepted Answer

When the model memorizes training data and then fails on new, unseen data

Answer

When the model performs equally well on both training and test data

Answer

When the model is too simple to learn any useful pattern

Answer

When training takes too long because the dataset is very large

Question 12

What is underfitting in a classification model, and what causes it?

Accepted Answer

When the model is too simple and fails to learn patterns in the data

Answer

When the model scores perfectly on all training examples

Answer

When the model is too complex and memorizes noise in the training data

Answer

When there are too many class labels for the model to handle

Question 13

What is a confusion matrix, and what information does it provide?

Accepted Answer

A table showing how often each actual class was predicted as each possible class

Answer

A table showing how confused the programmer is about the model's design

Answer

A chart displaying the training loss decreasing over time

Answer

A ranked list of the most important features in the model

Question 14

What is multiclass classification, and how does it differ from binary classification?

Accepted Answer

Sorting data into three or more predefined categories

Answer

Running several binary classifiers at the same time on different machines

Answer

Classification where there are only two possible outcome categories

Answer

Using multiple training datasets to build a single model

Question 15

What is precision in classification, and when is it especially important?

Accepted Answer

The fraction of predicted positives that are actually correct

Answer

The fraction of all actual positives that the model correctly identified

Answer

The overall accuracy of the model on the entire test set

Answer

The speed at which the model classifies new incoming data

Question 16

What is recall in classification, and when is high recall critical?

Accepted Answer

The fraction of all actual positives that the model correctly identified

Answer

The fraction of predicted positives that are actually correct

Answer

The percentage of training examples the model memorized exactly

Answer

A technique for refreshing and updating the training data

Question 17

What is cross-validation, and why is it more reliable than a single train/test split?

Accepted Answer

Training the model multiple times on different data splits to get a more reliable performance estimate

Answer

Checking that the code compiles and runs without any errors

Answer

Picking the best model by testing many models on the test set

Answer

Crossing out and removing noisy data before training begins

Question 18

What is class imbalance, and how can it be addressed?

Accepted Answer

When some classes have far more training examples than other classes

Answer

When all classes have exactly the same number of training examples

Answer

When some input features have missing values for certain examples

Answer

When labels are assigned incorrectly to some training examples

Question 19

What is a naive Bayes classifier, and why is it called 'naive'?

Accepted Answer

A probabilistic classifier based on Bayes' theorem that assumes all features are independent

Answer

A classifier that randomly selects different features for each prediction

Answer

A simple decision tree that only ever makes two branches

Answer

A classifier that always predicts the most common class in the training set

Question 20

What is feature engineering, and why can it matter more than the choice of algorithm?

Accepted Answer

Creating or transforming features from raw data to improve model performance

Answer

Removing all features from the dataset to keep the model as simple as possible

Answer

Letting the model automatically decide how many classes to predict

Answer

Writing feature names as variables in Python code

Question 21

What is an ensemble method, and how does Random Forest use it for classification?

Accepted Answer

Combining the predictions of many individual models to produce a better final result

Answer

A single very deep decision tree trained on the complete dataset

Answer

A method that randomly shuffles training labels before fitting the model

Answer

A database structure shaped like a forest for storing training data

Question 22

Give two real-world examples of classification problems and explain what makes them classification (not regression).

Accepted Answer

Identifying whether a handwritten digit is a 3 or a 7

Answer

Predicting exactly how many degrees warmer tomorrow will be

Answer

Calculating the average marks of students in a class

Answer

Finding the shortest driving route between two cities

Question 23

How does a classification model actually 'learn' from training data?

Accepted Answer

It adjusts its internal parameters to reduce prediction errors on labeled training examples

Answer

It memorizes every training example exactly, including all noise

Answer

It searches the internet to find the correct answer for each new example

Answer

It randomly picks a class for each example until it gets lucky and is right

Question 24

What is the key difference between classification and regression in machine learning?

Accepted Answer

Classification predicts discrete categories while regression predicts continuous numbers

Answer

Classification uses images as input; regression uses text as input

Answer

Regression always requires far more training data than classification

Answer

Classification is an unsupervised technique; regression is supervised

Question 25

What ethical problem can arise when a classifier is trained on biased data?

Accepted Answer

Biased training data can cause classifiers to discriminate unfairly against certain groups

Answer

Classification models always produce fair results as long as the code has no bugs

Answer

Ethical problems only arise when the training set has fewer than 1,000 examples

Answer

Ethics in classification only matters for image recognition systems, not text classifiers

Question 26

Despite its name, logistic regression is a classification algorithm. How does it work?

Accepted Answer

A classification algorithm that outputs class probabilities using a sigmoid probability curve

Answer

A regression algorithm used to predict continuous values like house prices

Answer

An unsupervised algorithm that groups data points without using labels

Answer

A method that always requires exactly two input features to work correctly

Question 27

What is a hyperparameter? Give an example from KNN and explain why choosing it carefully matters.

Accepted Answer

A setting chosen before training, like K in KNN, that controls how the model learns

Answer

A parameter the model learns automatically from the training data

Answer

The correct label assigned to a training example

Answer

A special type of feature extracted automatically from images

Question 28

What is the difference between a validation set and a test set in a classification project?

Accepted Answer

The validation set is used to tune hyperparameters; the test set gives the final honest evaluation

Answer

The validation set is used for training; the test set is not used at all

Answer

Both the validation set and test set are identical and serve the same purpose

Answer

The validation set is always larger than the training set

Question 29

How should you evaluate a multiclass classifier beyond reporting a single accuracy number?

Accepted Answer

Compute per-class precision, recall, and F1-score, and inspect a confusion matrix

Answer

Overall accuracy alone is sufficient for any multiclass problem

Answer

Only count errors on the rarest class, ignoring all others

Answer

Convert the problem to binary classification and re-evaluate from scratch

Question 30

What role do neural networks play in classification, and what makes them powerful?

Accepted Answer

Neural networks learn complex patterns through layered transformations and output class probabilities

Answer

Neural networks are only suitable for regression problems, not classification

Answer

Neural networks can produce accurate classifications without any training data

Answer

Neural networks are a simpler alternative to decision trees that always work better

Test yourself