ML Implementations in CSharp

Continuation of my ML Implementations in C++ repository, switched over to C# due to poor memory management in C++ when dealing with double pointers.

Old Repository

Requirements

• Microsoft .NET SDK 7.0
• Microsoft Visual Studio 2022

Running the code

From Source Code

Clone this Repository

git clone https://github.com/DarshK35/ml-implementation-cs

Open Project in Visual Studio

Select File > Open > Project/Solution in VS menu, then navigate to ml-implementation-cs.sln file cloned with the repo.

Run Code in Visual Studio

The code can be run using either of the following methods:

• Debug > Start without Debugging in VS Menu
• Ctrl + F5 shortcut key

Debugging can be left disabled for running this project, unless contributing to the code.

Release Version

Coming soon

Development Checklist

• Matrix Class
  • Overloaded arithmetic functions
  • Array-like indexing
• Linear Regression Class
  • Statistics-based Fitting function
  • Support for multiple independent variables
• Neural Network Class
  • Prediction Function
  • Actually Working Backpropogation Algorithm
  • Training algorithm
  • Multiple Loss function support (optional)
  • Multiple Activation Function support for Neural Network (optional)
• Linear SVM Classifier Class
  • Multiple kernel function support
• Add functions for Model parameter reviewing
• Menu-based Driver Code
• GPU Support (Maybe)

Notes

• Matrix class currently supports only matrix multiplication, Dot product is not supported yet
• Neural Network class is only useful for classification tasks, Regression tasks will be handled by separate activation function
• SVC class expects single output variable
• Kernel functions yet to be implemented