The impact of relevance feedback on web-based information retrieval for horizon scanning applications

Horizon scanning is being increasingly regarded as an instrument to support strategic decision making. It requires the systematic examination of data to identify potential threats and opportunities to improve resilience and decrease risk exposure. Horizon scanning may benefit from various retrieval techniques to augment the acquisition of data, though this involves a search for novel and emerging issues without knowing them beforehand. To optimise such a search, we propose the use of relevance feedback, which involves human interaction in the retrieval process so as to improve the results. As a proof-of-concept demonstration, we have carried out a horizon scanning exercise which showed that our utilisation of relevance feedback for horizon scanning applications was able to maintain the retrieval of relevant documents constant over the entire length of the experiment, without any reduction. This represents an improvement over previous studies where relevance feedback was not considered.