Exploratory Professional Search through Semantic Post-Analysis of Search Results

Abstract

Professional Search is usually a recall-oriented problem. For helping the user to get efficiently a concise overview, to quickly restrict the search space and to make sense of the results, in this article we present an exploratory strategy for professional search that is based on semantic post-analysis of the classical search results (of keyword based queries). The described strategy can exploit the metadata that are already available, as well as the results of textual clustering and entity mining that can be performed at query time. The outcome of this process (i.e. metadata, clusters and entities grouped in categories) complement the ranked list of results produced from the core search engine with useful information for the user. This extra information is useful not only for providing a concise overview of the search results, but also for supporting a faceted and session-based interaction scheme that allows the users to restrict their focus gradually and to explore other related information. To tackle the corresponding configuration requirements of this process, we show how one can exploit the (constantly evolving) Linked Data for specifying the entities of interest and for providing further information about the identified entities. In this article, apart from detailing the steps of this process, we present applications of this approach in the marine domain and in the domain of patent search.