# classification

## What is Classification in Machine Learning?

Classification in ML is when the model classifies a given input between available classes.

## Image Classification Problems

For Image Recognition/Image Classification, we usually use Neural Networks. More precisly when speaking of Images we use Convolutional Neural Networks.

## Value Classification Problems

What if we need to classify a value instead of an image?

Suppose we have a simple Yes or No example: we know the older you are, the more you have money to buy a house, or a product (maybe because of Salary or Savings).

In this case we want to predict if the person will be in the "Yes"(I buy it) or "No"(I don't) category.

We can approach the problem in more ways, to see which one fits better.

For value-classification problems, we usually do not use Neural Networks (you could), but we can apply:

- Logistic Regression - simply computes the average line-separator

- K-Nearest Neighbors (k-nn) - computation based on nearest neighbour

- Support Vector Machine (SVM) (kernel: 'rbf' (Gaussian), 'poly', 'sigmoid') - probabilistic methods

- Naive Bayes - probability

- Decision Tree and Random Forest - votation made by "trees" make the result

Thanks to SuperDataScience for their ML courses on Udemy!

https://www.superdatascience.com