Linear Regression From Scratch

A simple linear regression library written in C with a Python API.

___________________________________________________________________________/\\\________        
 __________________________________________________________________________\/\\\________       
  ____/\\\__/\\\___________________________________________________/\\\_____\/\\\________      
   ___\//\\\/\\\___/\\\\\\\\\\__/\\____/\\___/\\__/\\\____/\\\_____\/\\\_____\/\\\________     
    ____\//\\\\\___\//////////__\/\\\__/\\\\_/\\\_\///\\\/\\\/___/\\\\\\\\\\\_\/\\\\\\\\\__    
     _____\//\\\_____/\\\\\\\\\\_\//\\\/\\\\\/\\\____\///\\\/____\/////\\\///__\/\\\////\\\_   
      __/\\_/\\\_____\//////////___\//\\\\\/\\\\\______/\\\/\\\_______\/\\\_____\/\\\__\/\\\_  
       _\//\\\\/_____________________\//\\\\//\\\_____/\\\/\///\\\_____\///______\/\\\\\\\\\__ 
        __\////________________________\///__\///_____\///____\///________________\/////////___

Linear Regression

Linear regression is a method used to estimate a linear function (y = wx + b) that best fits a given dataset. This is done using a technique called gradient descent. Gradient descent is a more general optimization technique that is used also in other forms of machine learning. Simply put, gradient descent is an efficient way to find the best values for the w and b in y = wx + b.

For now the library is capable of univariate linear regression only, meaning the dataset should only consist of a pair of variables (x and y).

Using the Python API

To use the Python API simply import LinearRegressionModel from lib_linear_reg.py. Use the train() method to train the model, giving it two lists of integers (one for each variable). Once the model has been trained use the predict() method to create a prediction.

Check out example.py for an example of how to use the Python API.

Planned Features

• Precision
• Benchmarking
• Multivariate linear regression
• Plotting
• Parallelized gradient descent