What is Evidence in Decision Intelligence?
Evidence is the relevant data and logic that allows a user to validate a recommendation delivered by a decision intelligence platform. The evidence consists of insights that initially triggered the recommendation, as well as the supporting data used by the models and the DI platform. Often the evidence is in context with critical metrics and KPIs and includes the predicted impact of the recommendation expressed in the same terms that the KPI is measured. Evidence enables efficient and quick validation of the recommendation leading to rapid decision-making.
Diwo Explainable AI provides answers to why specific recommendations are generated by the system. With the plethora of machine learning models that are being used to make decisions today, it is important for business users to understand, validate and trust them as part of their decision-making process.
Diwo’s Evidence framework demystifies the black box of machine learning models at contextual granularity by:
- Explaining the different factors (with scores) that contributed to the predictions inference.
- Providing visualizations and highlights seen in the data used as input into the different models
- Statistics on metrics used by the model
- Insights into what data sources were used