Packt Publishing - Tensorflow Solutions for Text
English | Size: 416.76 MB
Use deep learning to make predictions on email
This course deals with the core topics in a highly practical manner with focus on code.
This volume introduces working with text, with a focus on the most plentiful source of text out there: email. Working with email text from your own Gmail account, you will build up a label predictor, similar in effect to the technology Google uses to power the Social and Promotions tabs. With this technique, you will be able to build your own email classification and automated workflow hooks.
What You Will Learn
* Fetch data from an email
* Use encoders to detect sample data
* Use recommenders to predict word similarity
* Predict output probabilities for data
* Build an automatic email server