Simple Machine Learning Application helping to organise your Email inbox
In the corporate environment, we usually receive from 50 to 200 emails a day and it's impossible to go through all of them. There is already a tool helping to manage that: Rules and Folders. In most of the cases, rules help you to filter traffic in a very primitive way like "move all the noreply notifications to some folder", but how we can improve this organization for emails which are really important for us and make it in a personalized manner. I would like to introduce an idea how to solve this problem in a more sophisticated way by introducing some machine learning techniques allowing to find a proper balance between important emails and others. Also, this talk will cover entire cycle of the development for Data product: from problem definition/idea to product app and integration.
Outline/structure of the Session
- Preamble and premises
- Problem definition
- Idea introduction
- Data preparation
- Classifier training
- Data product development
- Learnings and improvements
Data is everywhere. Just do it. How to build data product from scratch in your free time.
software engineers, processes managers, ml beginners
Classification problem and ML methods to solve it: LogReg, RandomForrest, etc