A Framework for Event-oriented Text Retrieval Based on Temporal Aspects: A Recent Review

Abstract

Event, as an important carrier for users to understand the world, has become a special retrieval object. In contrast to traditional text retrieval, Event-oriented Text Retrieval (ETR) can search events by utilizing events knowledge and using events as proxies for information needs. Accordingly, ETR has become the preferred way for users to obtain their interested events from massive web collections. Moreover, it also has already aroused considerable attention from scholars in recent years. However, the retrieval effectiveness of ETR is still subject to the effect of temporal aspects (i.e., temporal dynamics). Thus, in this review, we first analyze three major temporal components in the framework of ETR. After that, we provide a comprehensive overview of state-of-the-art approaches corresponding to such three components. Finally, we summary some ETR-related resources and pinpoint several potential research directions.