UcEF for Semantic IR

Abstract

At present, keyword-based techniques allow information retrieval (IR) but are unable to capture the conceptualizations in users' information needs and contents. The response to this has been semantic search computing with commendable success. Surprisingly, it is still difficult to evaluate Semantic IR (SIR) and understand the user contexts. The absence of a standardized cognitive user-centred evaluative paradigm (CUcEP) further exacerbates these challenges. This chapter provides the state-of-the-art on IR and SIR evaluation and a systematic review of contexts. Appropriate user-centred theories and the proposed evaluative framework with its integrated-context, web analytic conception, and related data analytic technique are presented. A descriptive approach is adopted, with the conclusion that multiple contexts are essential in SIR evaluation since “searching by meaning” is a multi-dimensional cognitive conception, hence the need to consider the impact of context dynamicity. Finally, the foregrounded semantic items will be applied to standardize the CUcEP in future.