摘要
有效的活动识别是智能辅助的关键。结合D-S证据和本体推理,提出一种互补结构的活动识别方法。该方法通过在证据理论和知识库之间建立对应关系形成互补,既解决了异构数据之间的知识共享,又对本体推理的规则冲突进行了处理,提高推理结果的准确性。通过复杂活动实例在原型系统中的应用,验证了该方法的可行性和有效性,并且能够有效地提高活动识别准确率。
The effective recognition of activity is the key to Intelligent Assistant. This paper proposes an activity recognition approach with complementary structure which combining D-S evidence theory and semantic reasoning. This approach creates the complementary through the correspondence between D-S theory and knowledge base,which can both resolve the knowledge sharing of heterogeneous data and the rule conflicts in ontology reasoning,and then improves the result of inference. Finally,the feasibility and effectiveness of the approach are verified through the application of complex activity in prototype system. Results show that it has better effect in improving the accuracy of activity recognition.
出处
《计算机工程与应用》
CSCD
北大核心
2016年第4期6-12,共7页
Computer Engineering and Applications
关键词
本体
本体推理
D-S证据理论
活动识别
ontology
ontology reasoning
D-S evidence theory
activity recognition