Objective: Design, develop and evaluate new methods for ranking query results that
- will exploit domain-specific heuristics (in particular in the clinical domain, most likely oncology), and
- will exploit “humans in the loop”
All our searches for information have a context and relate to some information or data. For example, clinicians want to extract targeted information that fits the context of a specific patient case (e.g. similar cases reported in literature or stored in a reference database, outcomes for that specific disease, best treatment options, etc.). Current solutions are only able to provide support for very simple questions and decisions, and are not able to fully address the increased complexity of clinical decision for example in the context of oncology.
We will investigate the notion of domain-specific expert ranking, and develop data models and heuristics to describe the notion of expertise, and to associate experts with specific domains. We will develop a system that will enable community-based expert feedback for content and for information sources. We incorporate these expert-driven domain-specific ranking heuristics into the query result ranking in WP3. We collect and provide such ranking to enable the extraction of targeted information relevant in a specific domain, for example to support clinically relevant scenarios. We will identify together with clinical experts concrete use cases, such as publication search in a specific clinical domain (most likely oncology), or clinical trial selec- tion based on expert feedback/recommendation.