Back to the Basics in Machine Learning” a Beginner’s Guide Quiz

Explore foundational concepts of machine learning, including key definitions, typical problem types, and core algorithm categories for beginners.

1. Defining Machine Learning

   Which statement best describes what machine learning is?

   1. A field of study that enables computers to learn from data without explicit programming
   2. A process of programming computers with fixed rules
   3. An approach that only uses decision trees for prediction
   4. A method to increase computer hardware speed

   Explanation: Machine learning allows computers to learn patterns from data and make predictions without being explicitly programmed for each task. Fixed rules are traditional programming, not machine learning. Decision trees are just one type of ML method; the field is broader. Increasing hardware speed is unrelated to the definition of machine learning.

2. Traditional vs. Machine Learning Approaches

   How does a traditional rule-based system differ from a machine learning system?

   1. Rule-based systems require manual updates, while machine learning adapts to new data automatically
   2. Rule-based systems perform better as datasets grow
   3. Both approaches learn from data automatically
   4. Machine learning systems cannot handle new data

   Explanation: Rule-based systems rely on manually crafted rules, making them inflexible to changes. Machine learning systems adapt as new data becomes available. Both approaches do not learn from data automatically—only machine learning does. Machine learning generally handles data growth better than static rule-based systems.

3. Types of Machine Learning Tasks

   Which example best illustrates a regression problem in machine learning?

   1. Grouping users by browsing behavior
   2. Identifying objects in an image
   3. Predicting the future sales volume of a product
   4. Classifying emails as spam or not spam

   Explanation: Regression problems involve predicting continuous numerical values, such as sales volume. Classifying emails is a classification problem, object identification is typically image classification or detection, and grouping users represents clustering, not regression.

4. Supervised vs. Unsupervised Learning

   What distinguishes supervised learning from unsupervised learning?

   1. Both require labeled data for training
   2. Supervised learning cannot be used for predictions
   3. Unsupervised learning is only used for number-based data
   4. Supervised learning uses labeled data, while unsupervised learning uses unlabeled data

   Explanation: Supervised learning algorithms learn from labeled data pairs, whereas unsupervised learning explores data without explicit labels. Both approaches can handle various data types, not just numbers. Supervised learning is primarily used for making predictions.

5. Instance-Based vs. Model-Based Learning

   Which key feature best characterizes instance-based learning?

   1. Instance-based learning only works with text data
   2. The system compares new examples directly to stored data instances
   3. Predictions are made without any reference to previous data
   4. The algorithm builds an explicit mathematical model beforehand

   Explanation: Instance-based learning methods store examples and make predictions by comparing new inputs to these stored instances. Model-based learning creates predictive models, so that option is incorrect. Predictions in instance-based learning always reference previous data. It is not limited to text; it can be applied to various data types.