Types of Machine Learning Algorithms
- Supervised Learning
- Unsupervised Learning
- Recommender Systems
- Reinforcement Learning
Supervised Learning
- Refers to algorithms that learn input-to-output mappings.
- Give learning algorithm examples to learn from that include the "output" label for a given input X.
- Eventually learns to take just the input alone without the output label and gives a reasonably accurate prediction or guess of the output.
- Learns from being given "right answers."
Examples of Supervised Learning
| Input (X) | Output (Y) | Application |
|---|---|---|
| Spam (0/1) | Spam Filtering | |
| Audio | Text Transcripts | Speech Recognition |
| English | Spanish, Chinese, etc | Machine Translation |
| Ad, User Info | User Click (0/1) | Online-Advertisments |
| Image, Radar Info | Position of other cars | Self-Driving Cars |
How does it learn? (Simplified)
- Split the data by 80% for training data and 20% for test data.
- Train the model with examples of Input (X) and Labels (Y) using the 80% training data.
- Use the 20% test data or unseen data, to predict or guess the output.
"Eventually learns to take just the input alone without the output label and gives a reasonably accurate prediction or guess of the output"
- Andrew Ng
Types of Supervised Learning Algorithm
1. Regression
- It is used to analyze the relationship between the independent variables and dependent variables.
- Predict a number from infinitely many possible numbers.
- Example: house prices (Y), size of the house (X)
- When you see a plot that uses linear regression, the Y always refers to the value we want to predict.
2. Classification
- We are trying to predict only a small number of possible outputs or categories.
- There are more than two possible outputs.
- Class/Category is the term that we use for the output.
- Predict categories, usually non-numeric.
- Find the boundary line that separates 0 and 1.
- Breast Cancer Detection
[0: benign, 1: malignant] - Examples: Cat or Dog, Benign or Malignant
Top comments (0)