基于民航附加服务本体的个性化协同过滤算法

Personalized collaborative filtering algorithm based on auxiliary service ontology of civil aviation

在构建完成的民航附加服务本体的基础上,针对传统协同过滤算法存在数据稀疏性和无法考虑项目自身语义关系等缺点,提出了基于本体的个性化协同过滤算法。该方法采用基于规则的推荐方法在Jena的基础上进行推理,然后利用推理结果更新本体,完善了本体中旅客的兴趣评分信息,降低了旅客兴趣评分稀疏性。接着引入本体语义相似度计算,将语义相似度与协同过滤算法相似度结合,加入了项目之间语义相似度的计算,提高了相似度计算的准确度。通过实验证明,改进的推荐技术很好地提高了推荐的准确性。

Personalized ontology -based collaborative filtering algorithm is proposed, aiming at the short -comings oftraditional collaborative filtering algorithm such as data -sparsity and lack of reflection to the semanticrelationship of item itself. The method adopts Jena inference engine and constructs association rules based onontology inference rules, and inference results are used to update ontology and complete the interest ratinginformation of tourist to reduce the sparsity of tourist interest rating. Then the ontology semantic similarity isintroduced to compute and combine the similarity of semantics and the similarity of collaborative filteringalgorithm, improving the accuracy of similarity calculation after adding semantic similarity computation betweenitems. Experiments prove that the hybrid recommendation technology effectively improves its accuracy.