https://github.com/anuraganalog/machine-learning

Machine Learning with some codes.
https://github.com/anuraganalog/machine-learning

cost-funtion datasets from gradient-descent linear-regression machine-learning metrics ml neural-networks plotting python3 scratch statistics testing training written

Last synced: 2 days ago
JSON representation

Machine Learning with some codes.

• Host: GitHub
• URL: https://github.com/anuraganalog/machine-learning
• Owner: AnuragAnalog
• License: mit
• Created: 2020-01-01T02:04:23.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2023-07-06T21:56:16.000Z (over 1 year ago)
• Last Synced: 2024-11-18T07:41:41.702Z (about 1 month ago)
• Topics: cost-funtion, datasets, from, gradient-descent, linear-regression, machine-learning, metrics, ml, neural-networks, plotting, python3, scratch, statistics, testing, training, written
• Language: Python
• Homepage:
• Size: 184 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Machine Learning

## Prerequsities

* Probability

* Statistics

* Linear Algebra

* Calculus

To get a basic understanding on the statistics, you can try [this](https://www.hackerrank.com/domains/tutorials/10-days-of-statistics) course in HackerRank.

And, [here](https://github.com/AnuragAnalog/hackerrank/tree/master/tutorial-10-days-of-statistics) is the solutions.

## Datasets

* Random Dataset for Linear regression

## Loss functions

### Mean Squared Error(MSE)

It is the average squared error between the actual and the predicted observations, it is also called as Quadratic Loss, L2 Loss.

This method penalizes the values which have high error.

### Mean Absolute Error(MAE)

It is the average absolute error between the actual and the predicted observations, it is also called as L1 Loss

This method is more sensitive to outliers than MSE, this also needs more complex techniques to compute the gradients.

### Mean Bais Error

It is just the difference between the actual and the predicted value, we should take a bit care when dealing with positive and negative bais, because they can make the total bais less.