My Progress

1. IntuitionLinear Regression is not a good method for Classification Problem.Why?One outlier on the right changes the linear regression function. It moves decision boundary dramatically. 2. Logistic Regression2.1 FormulaLogistic function allows the function to be curved unlike the linear regression. This is a sigmoid function aka logistic function. x-axis represents the number z. It outputs a v..

1. Checking Gradient descent for convergence How to check if Gradient descent is working well? 1.1 Graph This graph is called Learning curve. As the iteration increases, the minimum cost should decrease. If the graph remains constant after enough iterations, we call that it has converged 1.2 Epsilon / Automatic convergence test Let epsilon be 0.001 If the cost function decreaes by epsilon in one..

1. Intuition House example: Lets say we are predicting the price of a house based on its size and number of bed rooms. Number for size of a house is relatively larger than the number of bed rooms. This large difference between numbers will make it hard to accurately predict the price. How is this related to gradient descent? Since the number for size is large, it takes smaller w value to make a ..