Feedback obtained: "Almost perfect. The hessian term is always using L2 regularization."
Criteria: Create a custom, one-versus-all logistic regression classifier using numpy and scipy to optimize. Use object oriented conventions identical to scikit-learn. You should start with the template developed by the instructor in the course. You should add the following functionality to the logistic regression classifier:
Ability to choose optimization technique when class is instantiated: either steepest descent, stochastic gradient descent, or Newton's method.
Update the gradient calculation to include a customizable regularization term (either using no regularization, L1 regularization, L2 regularization, or both L1 and L2 regularization). Associate a cost with the regularization term, "C", that can be adjusted when the class is instantiated
Feedback obtained: "Almost perfect. The hessian term is always using L2 regularization."
Criteria: Create a custom, one-versus-all logistic regression classifier using numpy and scipy to optimize. Use object oriented conventions identical to scikit-learn. You should start with the template developed by the instructor in the course. You should add the following functionality to the logistic regression classifier: Ability to choose optimization technique when class is instantiated: either steepest descent, stochastic gradient descent, or Newton's method. Update the gradient calculation to include a customizable regularization term (either using no regularization, L1 regularization, L2 regularization, or both L1 and L2 regularization). Associate a cost with the regularization term, "C", that can be adjusted when the class is instantiated