Applied Machine Learning

You name it!

In this course, we cover some of the most popular and practical machine learning algorithms that are used in the industry. An industry use case comes with each algorithm to ensure students understand how to apply those models in real world.

• Linear Regression (lasso/ridge)
• Time Series Forecasting
• Logistic Regression
• Decision Trees and Random Forests
• Gradient Boosting Trees (Xgboost, Catboost)
• Factorization Machines
• Recommender Systems (Collaborative Filtering)
• Neural Networks
• Unsupervised models: K-Means and Hierarchical Clustering
• Dimension reduction with PCA and feature screeningdf

We teach students popular frameworks such as Scikit-learn, Gensim, NLTK, and Keras. AutoML frameworks such as H2O will also be covered in the lab.

• Supervised learning models using Scikit-learn
• Unsupervised learning such as clustering using Scikit-learn
• Text processing using NLTK
• Topic modeling and word2vec using Gensim
• Neural network models using Keras
• AutoML using H2O
• Parameter Tuning with Gridsearch, RandomSearch and Bayesian Optimization

Feature preprocessing ability separates the good data scientists from the bad

In real-world data scientist jobs, more efforts are spent on data munging and engineering features. Feature screening and selection also make sure the models are not overfitting. Most textbooks and online courses don't cover these important topics in detail. You will learn many tips and tricks in this course regarding feature engineering, screening, and selection

• Feature screening methods
• Feature selection methods
• Featuretools
• Feature Transformers
• Categorical/Numerical encodings
• Feature Union
• Pipelines

Real-world use cases taught by real-life data science practitioners

• ML in retail
• ML for marketing response modeling
• ML for text mining, sentiment analysis
• ML for credit risk
• ML for offer optimization
• ML for business forecasting
• ML for personalized recommendations

Practice makes perfect. Students will come up with their own ideas and work on two capstone projects in this course so they can use it for job search. Original project ideas with strong motivation are always suggested.

• Project 1: Building supervised classification models
• Project 2: Building supervised regression models
• Project 3: Unsupervised models (optional)