⚠️ Work in Progress. If the users do not trust a model or a prediction, they will not use it.
Visual evaluation of machine learning model
UX Designer | June 2020
Problem
The classification decisions made by machine learning models are difficult to understand. For Business experts who train these models, it’s essential to evaluate the model before deploying it. To make this decision, users need to be confident that the model will perform well on real-world data, according to the metrics. Machine learning engineers use a lot of metrics to evaluate a model. How to present these metrics in a meaningful way, so users with no special knowledges in machine learning model could still make a right decision?
Product
Content Analyzer is an enterprise web application to classify documents. It has four-step workflow:
Manage data
Train model
Evaluate model
Deploy model
I worked non Evaluate model part
Persona
It architect, Busiiness expert with basic knowledge about training machine learning model.
Before
After user trained a model in IBM Content Analyzer user will see this page to evaluate model before deplyment.
