Background Materials

Leaf classification is a perfect area for the application of CNN$.$ This week, you will compete in the Kaggle.com Leaf ClassificationLinks to an external site. competition.

RECOMMENDATION: the data set here is large, so you might consider conducting exploratory data analysis and building initial models on a smaller subsample prior to applying it to the full dataset. Eventually, you will need to build your models on the entirety, though.

Management$/$Research Question

In layman$$s terms, what is the management$/$research question of interest, and why would anyone care?

Requirements

Conduct your analysis using a cross$-$validation design.

Conduct EDA.

Build at least three CNN models based on hyperparameter tuning.

Evaluate goodness of fit metrics.

Build multi$-$class ROC and Precision $/$ Recall graphs.

Once you have your best$-$performing models, classify the test data and submit it to Kaggle. Provide your Kaggle.com user name and screen snapshots of your scores.

Discuss your model$$s performance.

Python Programming

All programming will be done in Python. A style guide is hereLinks to an external site..