Abstract and Introduction

Points

Clearly summarizes the motivation of the project and main results. It is clearly explained why the problem is important and why it makes sense to use a machine learning approach over a traditional method.

10

Data

Data is clearly explained with key summary statistics. Data was   plotted to support the project motivation. Domain knowledge was used to create the features and clean the data.

10

Methodology

Project incorporated most methods from the class including lasso, ridge, principal component regression, random forest, boosting,   linear/logistic regression, support vector machines, KNN. The      validation set approach or CV was used to tune parameters. Deep learning can be used to replace some methods above.

20

The machine learning methods are clearly and intuitively described (i.e., researcher is not simply using ML packages without deeply    understanding the methods). It is clearly described how you            managed to solve the problem using machine learning tools, and    why these methods can outperform the traditional benchmark.        Proper methods for model evaluation and comparison were used.

20