Logistic Regression with Python and Numpy

Overview

Class Central Tips

Welcome to this project-based course on Logistic with NumPy and Python. In this project, you will do all the machine learning without using any of the popular machine learning libraries such as scikit-learn and statsmodels. The aim of this project and is to implement all the machinery, including gradient descent, cost function, and logistic regression, of the various learning algorithms yourself, so you have a deeper understanding of the fundamentals. By the time you complete this project, you will be able to build a logistic regression model using Python and NumPy, conduct basic exploratory data analysis, and implement gradient descent from scratch. The prerequisites for this project are prior programming experience in Python and a basic understanding of machine learning theory. This course runs on Coursera's hands-on project platform called Rhyme. On Rhyme, you do projects in a hands-on manner in your browser. You will get instant access to pre-configured cloud desktops containing all of the software and data you need for the project. Everything is already set up directly in your internet browser so you can just focus on learning. For this project, you’ll get instant access to a cloud desktop with Python, Jupyter, NumPy, and Seaborn pre-installed.

Syllabus

Deep Learning Fundamentals: Logistic Regression

Welcome to this project-based course on Logistic Regression. In this 2-hour long project-based course, you will learn how to implement Logistic Regression using Python and Numpy. Logistic Regression is an important fundamental concept in Deep Learning, and even though popular machine learning frameworks have implementations of logistic regression available, learning to implement it on your own will enable you to understand the mechanics of optimization algorithm and the training and validation process. By the end of this course, you would create and train a logistic model that will be able to predict if a given image is of hand-written digit zero or of hand-written digit one. The model will be able to distinguish between images or zeros and ones, and it will do that with a very high accuracy. Not only that, your implementation of the logistic model will also be able to solve any generic binary classiﬁcation problem. You will still have to train model instances on speciﬁc datasets of course, but you won’t have to change the implementation and it will be re-usable. The dataset for images of hand written digits comes from the popular MNIST dataset. This data set consists of images for the 10 hand-written digits (from 0 to 9), but since we are implementing logistic regression, and are looking to solve binary classiﬁcation problems, we will work with examples of hand written zeros and hand written ones and we will ignore examples of rest of the digits.