Roles of event actors and sentiment holders in identifying event-sentiment association

Abstract

In this paper, we study the roles of event actors and sentiment holders from the perspective of event sentiment relations within the TimeML framework. The proposed algorithm is bootstrapping in nature that identifies the association between the event and sentiment expressions. There are two basic steps of the algorithm and they deal with lexical keyword spotting and co-reference resolution. We consider the associations between the event and sentiment expressions that are in the same or different text segments. Guided by the classical definitions of events in the TempEval-2 shared task, a manual evaluation is attempted to distinguish the sentiment events from the factual events and the agreement was satisfactory. In order to computationally estimate the different sentiments associated with different events, the knowledge of event actors and sentiment holders is introduced. To identify the roles between the event actors and sentiment holders, appropriate method is proposed. From the experiments, it is observed that the lexical equivalence between event and sentiment expressions easily identifies the similar entities that are both responsible for the event actors and sentiment holders. If the event and sentiment expressions occupy different text segments, the identification of their corresponding event actors and sentiment holders needs the knowledge of parsed-dependency relations, named entities along with the anaphors. The manual evaluation produces satisfactory results on the test documents of the TempEval-2 shared task in case of identifying the many to many associations between the event actors and sentiment holders for a specific event. © 2012 Springer-Verlag.