Lecture
Purpose of Machine Learning Models
The purpose of machine learning models is to learn from data to solve certain problems.
For example, they can be utilized to distinguish spam emails or to predict housing prices.
The types of problems that machine learning models address generally fall into two categories:
-
Predicting specific categories (classes)
Classification -
Predicting continuous numerical values
Regression
Differences Between Classification and Regression
The differences between classification and regression problems are as follows:
| Category | Classification | Regression |
|---|---|---|
| Output | Specific class (e.g., Spam/Normal) | Continuous numeric value (e.g., $100,000) |
| Examples | Cat vs. Dog | Height prediction in inches |
| Objective | Grouping data | Predict numerical values |
When creating a machine learning model, it's important first to determine whether you're dealing with a classification or regression problem.
In the next lesson, we will explore classification in more detail.
Lessons in this chapter · How Machine Learning Models Work
- 1. How AI Models Learn
- 2. Training Dataset: Learning Patterns
- 3. Validation Dataset: Tuning the Model
- 4. Test Dataset: Final Performance Check
- 5. Multiple Choice Quiz
- 6. Hyperparameters: Key to Model Performance
- 7. Learning Rate: Controlling Training Speed
- 8. Batch Size: How Much Data to Learn at Once
- 9. Epochs: Number of Times to Train
- 10. Overfitting: A Closer Look
- 11. Underfitting: A Closer Look
- 12. Fill-in-the-Blank Quiz
- 13. What Is the Purpose of ML Models?
- 14. Classification Models: Grouping Data
- 15. Accuracy: Measuring Prediction Quality
- 16. Precision: Measuring Correctness of Positive Predictions
- 17. Recall: Measuring Coverage of Positive Predictions
- 18. F1-Score: Balancing Precision and Recall
- 19. Multiple Choice Quiz
- 20. Regression Models: Predicting Continuous Values
- 21. Mean Squared Error (MSE)
- 22. Mean Absolute Error (MAE)
- 23. R-Squared (R²)
- 24. Fill-in-the-Blank Quiz
Quiz
0 / 1
What is the most appropriate word for the blank below?
A regression problem involves predicting values.
continuous
discrete
integer
character
Lecture
AI Tutor
Design
Upload
Notes
Favorites
Help