Multiobjective Sensor Ontology Matching Technique with User Preference Metrics

Due to the problem of data heterogeneity in semantic sensor networks, communications among different network applications are seriously hampered. Although ontology is regarded as state-of-the-art knowledge model for exchanging information, there also exists between ontologies. Ontology matching an effective method deal with problem, whose kernel technique similarity measure. How integrate measures determine alignment high quality users preferences a challenging problem. To face this challenge, our work, Multiobjective Evolutionary Algorithm (MOEA) used determining nondominated solutions. In particular, evaluating metric on alignment’s proposed, which takes into consideration user’s and do not need use Reference Alignment (RA) beforehand; optimization constructed define formally, selection operator presented, can make MOEA uniformly improve solution’s objectives. experiment, benchmark from Evaluation Initiative (OAEI) real ontologies domain test performance approach, experimental results show validity approach.