Acquiring knowledge from categories using semantic associations

Abstract

Currently, there are several RDF (Resource Description Framework) knowledge bases that store facts about entities and community-generated categories of entities. These two types of knowledge may have strong associations

for example, entities categorized in "People from Korea" have a high probability of being born in Korea. Some of such associations can be used for predicting new facts about entities. In this paper, we propose a prediction system that predicts new facts from categories of entities. First, the proposed system uses novel association rule mining (ARM) approach that effectively mines rules that encode associations between facts and categories of entities in RDF knowledge bases. Our extensive experiments show that our novel ARM approach outperforms the state-of-the-art ARM approaches in terms of the prediction quality and coverage of the mined rules. After rules are mined, the proposed system ranks and groups the mined rules based on their predictability by our novel semantic confidence measure calculated with various semantic resources such as WordNet and embedded word vectors. The experiments show that our novel confidence measure outperforms the widely used standard measure in terms of discriminating the predictability of the mined rules. The proposed prediction system selects predictive rules from the mined rules ranked and grouped by their predictability, and then use them to predicts new facts of the high precision from categories of entities. The experiments show that the results of the proposed prediction system are fairly comparable to the results of the state-of-the-art prediction system, but with the high coverage of relations.