Spam Detection

Building a neural network to classify emails as spam, as well as "traditional" competitor models. Models are built in R/models.qmd. The rendered notebook is also attached as HTML.

Neural Network

Sequential model (multilayer perceptron) built with keras/tensorflow. Key facts: Adam optimizer, weight decay (L2-regularization), learning rate scheduling, early stopping. This is the architecture:

Benchmark against traditional models

Test metrics & performance

Model	Precision	Accuracy	F1
Neural Network	0.932	0.939	0.922
Random Forest	0.935	0.942	0.926