Upgrade Linear Regression with Fully Vectorized Implementation and add a naive implementation for beginners

[somrita-banerjee]

Feature description

I noticed that the current linear_regression.py
Implementation is partially vectorized but still uses Python loops and manual parsing in some places. I think for a learning platform, we should include both naive and vectorized algorithms.

[pushkar-hue]

Hello,

My friend @i3hz and I would like to collaborate on this issue. We agree that adding both naive and vectorized versions is a valuable first step for the library's educational goal.

As a potential future enhancement, once the initial task is complete, we would also be interested in exploring a CUDA implementation to demonstrate GPU acceleration for enhanced runtime. We could tackle that in a separate PR if the maintainers are interested.

Could you please assign this to one of us to get started? We will coordinate the work.

Thanks!

[somrita-banerjee]

I am already working on this

[i3hz]

As mentioned before me and @pushkar-hue are also working on implementing gpu support via cuda so if you'd like we can work together on the issue . Or else we can work on the cuda implementation seperately

[somrita-banerjee]

Thanks for the suggestion! I’m already working on this issue, so I’ll continue with the implementation. You’re welcome to work on a separate CUDA implementation — that way, both efforts can complement each other.