Machine Learning Exam - Easy and Complete Answers
SECTION - A (2 Marks Each)
1. Types of Supervised Learning:
- Classification: Predicts categories.
- Regression: Predicts continuous values.
2. Dimensionality Reduction:
- Reducing input features while keeping important information.
- Example: PCA (Principal Component Analysis).
3. Classification:
- Predicting categories or labels.
- Example: Spam or Not Spam email.
4. Clustering:
- Grouping data based on similarity.
- Example: Customer segmentation.
5. Labelled vs Unlabelled Data:
- Labelled: Has output labels, used in supervised learning.
- Unlabelled: No output labels, used in unsupervised learning.
6. Python in ML:
- Libraries like Scikit-learn, TensorFlow, Pandas.
- Easy data handling and model building.
SECTION - B (5 Marks Each)
7. Scikit-learn:
- Python ML library.
- Features: Easy to use, supports classification, regression, clustering.
- Built on NumPy, SciPy, Matplotlib.
8. Supervised vs Unsupervised Learning:
- Supervised: Uses labeled data, output known.
- Unsupervised: Uses unlabeled data, output unknown.
9. K-means Limitations:
- Must set number of clusters.
- Sensitive to outliers.
- Poor with non-spherical clusters.
- Depends on initial selection.
10. ML Applications:
- Spam filtering, Recommendations, Face/Speech recognition, Self-driving cars, Fraud detection.
11. Naive Bayes:
- Based on Bayes' Theorem.
- Assumes features are independent.
- Used in text classification.
- Fast and efficient.
12. Data Preparation Steps:
- Data Collection
- Data Cleaning
- Data Transformation
- Feature Selection
- Data Splitting
