Context-based health information retrieval

in HIR applications. We intent to characterize this behav- ior through the application of questionnaires to a sample of health professionals, interviews to a smaller set of pro- fessionals and, if logs are available, through the analysis of searches made on general and specialized search engines. The next step is to propose an information retrieval framework that defines what, how and where will the pre- viously identified contextual features be used in the IR sys- tem. Another necessary step involves the development of au- tomatic methods to collect contextual features (e.g.: to identify the searcher's expertise or a search's scenario). This may done through the analysis of the searcher's or group's characteristics, past interaction processes, documents' char- acteristics or even by exploiting temporal properties like fre- quencies and length of searches. This data may be in log files and it can be acquired from other systems (e.g.: searches' history, user's bookmarks, email or o"ce applications). This framework will be used to build a prototype, based on standard IR models, that integrates methods for the col- lection of contextual features.