Semantic XAI for contextualized demand forecasting explanations

Explanations can protect privacy while remaining useful: The system generates understandable forecasts without exposing sensitive details about how the underlying demand forecasting models work, demonstrating that transparency and privacy protection aren't mutually exclusive.

Context matters for AI explanations: By combining forecast information with related media events and external data through a knowledge graph, the explanations provide real-world context that helps users understand why a forecast was made, not just what was predicted.

Higher-level concepts improve usability: Rather than explaining forecasts using raw technical features, the system uses abstract concepts from a knowledge graph, making explanations more accessible to non-technical stakeholders who need to act on the forecasts.

Actionable insights increase practical value: The architecture specifically emphasizes dimensions where users can actually take action, moving beyond pure explanation toward decision support that helps businesses respond to forecasts.

Open resources advance the field: By making their ontology and datasets publicly available, the researchers enable other organizations to build similar explainable systems for demand forecasting and related domains.