Semantic Relationships Extraction for Entities in WIS

Web Integration System (WIS) provides abundant structured information about entities in a domain. Interrelationships for entities in WIS are valuable for further analysis and decision-making. It is not rare that an entity pair has more than one semantic relationship. However, existing researches on relation extraction ignore this situation and they assume that one entity pair has only one semantic relationship. This paper focuses on mining multi-semantic relationships for a giving entity in WIS. We first extract related entities and corresponding contexts from web texts, then propose a clustering algorithm to cluster the related entities into different subsets, where each subset represents a semantic relationship to the entity. Finally we adjust the result clusters by merging semantic similar clusters together. The highlight of the algorithm is that we cluster the entities based on every single context instead of the overall contexts related. We evaluate our method by comparing it with the state-of-the-art approach using real-world dataset generated by search engine. The results show that the proposed approach is efficient in mining multi-semantic relationships for the giving entity from WIS.