by mostly hand code to practice understanding & develope the optimize skill application and algorithm's core idea
Also practice using software design concept to build package
1.logist_reg_gradient_by_hand.py:logistic regression use gradient descent to optimize and also show the decision boundary
*can get top 29% in kaggle titanic data set tutorial by only using this hand code class to predict
2.back_propagation_ann.py: artificial neural network implement backpropagate algorithm
*first layer is tanh , second layer is softmax
3.ada_grad.py:adagrad optimization function