I am currently trying to analyse categorical data in an unsupervised manner to perform anomaly detection. As part of this, I have recently started looking into a couple of items like autoencoders and different strategies to convert my categorical data into values that autoencoders accept as input, especially with the processing being less resource intensive, capable of running on commodity hardware. Does your book help identify practices/use of tensorflow 2.0 from a use-case based perspective like this and how in-depth does the content go to outline some of the amazing resources that tensorflow has on offer?
But hopefully with the knowledge provided here, you'd be able to generalize that to your task.
PhD | Senior Data Scientist | AI/ML Educator
Rahul Dayal Sharma
posted 1 month ago
It would be quite helpful if I can identify some usable methods to probably understand how to convert categorical data into something more meaningful for autoencoders to process. The listed use cases sound interesting, especially the Sentiment analysis one.