A single layer neural network with two nodes and a output layer has been built from scratch in python. Since it contains 3 nodes(including the output node), I have initialised 6 weights and 3 biases. The gradient for each parameters has been learnt by the model through the iterative process. Prediction is done with those learnt parameters. I have declared everything under Class for code reuseability. The above classification model has a test accuracy of 90%. The libraries used are Numpy, Pandas, Matplotlib and Scikit-learn.
I also explained this code in my youtube channel named "Deep Matrix". you can see this video here.