Insights
Unleashing the power of feature importance analysis with Insights
What are "Insights"?
Insights represent an advanced feature importance calculator designed to comprehensively understand your model's performance on a training set. This will enable you to separate genuine from high-risk customers. With this cutting-edge functionality, you can now unravel the importance of each signal, gaining full explainability and valuable insights into the factors that drive your model's predictions.
By collecting a wide range of data the feature will allow you to easily identify trustworthy customers. At the same time, you can use Insights to red-flag suspicious activity, thus isolating potential fraudsters.
How does it work?
βInsightsβ employs sophisticated algorithms to calculate scores for each input feature based on their impact on the model's predictions. A higher score represents a greater influence within the model, helping you understand, prioritize, and optimize your rule sets for enhanced model performance.

Begin by selecting the "Insights" tab on the Trustull dashboard. Upload your training set effortlessly, defining "OK" and "KO" users with a simple click, and, watch as the platform calculates feature importance scores without any coding. Explore the results instantly, gaining valuable insights to optimize, debug, and enhance decision-making for your risk modeling projects.

It's a hassle-free process, putting the power of model interpretation at your fingertips!
Key Benefits:
- Transparency: Gain transparency into your model's decision-making process by understanding the significance of each signal.
- Optimization: Identify and focus on the most influential features to optimize your model and enhance its predictive accuracy.
- Decision support: Make more informed decisions by leveraging insights into the factors driving your model's predictions.
- Model debugging: Pinpoint potential issues or biases by closely examining the importance scores of different features.
- Enhanced explainability: Communicate model behavior more effectively to stakeholders and build trust in your risk decision-making.
Updated 9 months ago