**Classifiers** are algorithms that map input data to a class (or category).
Optical Character Recognition is the quintessential example of a classifier. However, classifiers can be simpler.
A classifier that separates inputs into one of two results is called a **binary classifier**. A pregnancy test is a binary classifier. There are [[Binary Classifier Evaluation]] metrics that are worth learning about.
# Types
- **Decision Tree** - typically binary trees representing choices that work their way down, ideally with leaves that resolve to a single class
- **Support Vector Machine** - basically where you can plot out values on a [[Scatter Chart]] then draw lines to separate the different classes
- **Neural Net** - the basis of artificial intelligence, where inputs go through successive layers of *logic units* which have their weights adjusted through **back propagation** to better classify stuff
- The famous "Perceptron" is the first of these
- These use [[Sigmoid]] functions under the hood
****
# More
## Source
- grad school
## Related
-