Authors: Jinchuan Chen, Jingjing Zhang, Ju Fan, Xuan Wang
Tags: 2018, conceptual modeling
Rules or constraints can be used to clean a knowledge base, or find new facts which should have been included. Recently there are many efforts on automatically mining rules from large scale knowledge bases. However, these rules usually contain no constants. In practice, we often need some detailed rules, for example, rules restricted to a special country or a special profession. One major challenge of appending constants lies in that there are large amount of constants, each of which can generate a new rule. Moreover, we have to choose appropriate granularity in order to trade off between the applicability and precision (or support and confidence in traditional rule mining terminology). In this paper, we propose a Spark based solution to mine rules with constants, a taxonomy based approach to control the granularity, and several techniques to improve the efficiency. We also conduct extensive experiments to evaluate the efficiency and effectiveness of our solution with comparison with the state of the art works.Read the full paper here: https://link-springer-com.proxy2.hec.ca/chapter/10.1007/978-3-030-00847-5_38
