In this project, we explore deep learning with Keras using two different datasets and tasks.
-
Modify the convnet example from the chapter to load and process the Fashion-MNIST dataset instead of the MNIST dataset. This involves importing the correct module, loading the Fashion-MNIST data, and running the model with these images and labels.
-
Evaluate how well the model performs on Fashion-MNIST compared to MNIST. Analyze the model's accuracy and performance.
-
Compare the training times for both datasets to assess any differences in training speed.
-
Research SPAM email detection using deep learning and Keras. Utilize the Spambase Dataset from the UCI Machine Learning Repository.
-
Implement a deep-learning binary-classification model that predicts whether emails are SPAM or not using recurrent-neural-network techniques learned in Chapter 16.
-
Investigate other SPAM email datasets and attempt to use them with your model to further improve accuracy.