Data Science with Python: Machine Learning

This 20-hour Machine Learning with Python course covers all the basic machine learning methods and Python modules (especially Scikit-Learn) for implementing them. 

The five sessions cover: 

• simple and multiple Linear regressions;
• classification methods including logistic regression, discriminant analysis and naive bayes, support vector machines (SVMs) and tree based methods; 
• cross-validation and feature selection; 
• regularization; 
• principal component analysis (PCA) and clustering algorithms. 

After successfully completing of this course, you will be able to explain the principles of machine learning algorithms and implement these methods to analyze complex datasets and make predictions in Python.

Prerequisites

• Knowledge of Python programming
• Able to munge, analyze, and visualize data in Python

Syllabus

Unit 1: Introduction and Regression

• What is Machine Learning
• Simple Linear Regression
• Multiple Linear Regression
• Numpy/Scikit-Learn Lab

Unit 2: Classification I

• Logistic Regression
• Discriminant Analysis
• Naive Bayes
• Supervised Learning Lab

Unit 3: Resampling and Model Selection

• Cross-Validation
• Bootstrap
• Feature Selection
• Model Selection and Regularization lab

Unit 4: Classification II

• Support Vector Machines
• Decision Trees
• Bagging and Random Forests
• Decision Tree and SVM Lab

Unit 5: Unsupervised Learning

• Principal Component Analysis
• Kmeans and Hierarchical Clustering
• PCA and Clustering Lab

Final Project

After 20 hours of structured lectures, students are encouraged to work on an exploratory data analysis project based on their own interests. A project presentation demo will be arranged afterwards.

Recommended Readings

• An Introduction to Statistical Learning, by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani
• Applied Predictive Modeling, by Max Kuhn and Kjell Johnson
• Machine Learning for Hackers, by Drew Conway, John White