The assignment

Find an interesting data set and construct a machine learning model from the data set. You can either do a regression problem or a classification problem.

Additional requirements

In the course of working with your data set try to demonstrate at least one idea that is covered in the textbook but was not covered well in the lectures. Examples of things to demonstrate include

• Using grid search to optimize the choice of model hyperparameters
• Constructing an ensemble of different models
• Using the k-nearest neighbors algorithm (KNN)
• Using a non-linear kernel with PCA or a support vector machine
• Demonstrating some other dimension-reduction algorithm, such as LLE or Isomap
• Using online learning to refine a model as new data arrives

Where to look for data sets

If you need help finding a data set to work with, there are several sources available online where you can find example data to work with or links to data sources. Here is a list of some of the more common sources.

• UCI Machine Learning Repository
• mldata.org Repository
• Data is Plural
• bigml Blog List of Public Data Sources

Group project

This final project is a group project. Please form groups of 2-3 people each to work together on your project. Each group should prepare a Jupyter notebook to turn in, and each group will give a brief demonstration of their work in class on Tuesday, March 6.