基于个性化特征的无公害农产品目录智能服务系统

Intelligent service system of pollution-free agricultural products catalog based on personalized features

从促进无公害农产品流通,提高认证产品信息服务质量出发,该文以无公害农产品目录为研究对象,以实现认证产品信息的个性化检索与推荐为目标,进行产品目录智能服务的关键技术研究。以产品概念、属性和语义关系为主体,设计产品目录本体模型,建立产品目录领域本体,形成无公害农产品电子化目录;基于产品目录领域本体,从本体概念层面进行用户兴趣特征建模,采用语义清晰度与用户兴趣度的综合值设置概念权值,建立个性化产品目录加权本体模型,并设计个性化产品目录加权本体建立算法,实现用户兴趣的动态、准确和全面描述;通过个体过滤、综合过滤、概念映射、语义查询扩展和查询结果优化,构建个性化产品目录信息推荐与检索算法,建立产品目录智能服务系统。在检索结果用户满意度对比试验中,相比于基于领域本体检索、基于语义相似度优化检索和基于个性化本体优化检索3种方式,分别提高了31%、21%和14%,结果表明可在一定程度上提高无公害农产品目录信息检索推荐的质量和满意度,满足个性化的需求,同时也可为无公害农产品推广、品牌化销售和电子化交易提供技术支撑,具有较好的应用前景。

For promoting pollution-free agricultural products＇ circulation and improving the service quality of the certified product information, to realize the personalized information retrieval and recommendation of the certified product information, this paper carried out key technology research of the pollution-free agricultural products catalog intelligence services. Employing the product concept, properties, and semantic relations as the principal part, the ontology model of a product catalogue was designed. Moreover, the product directory domain ontology was also set up to form the pollution-free agricultural product electronic catalog. Based on the products catalogue domain ontology, the users＇ preference feature model was conducted from the ontology concept level. The personalized product catalog weighted ontology model was established based on the concept of weights set by the comprehensive value of the semantic resolution and users＇ degree of interest. Furthermore, the personalized product catalog weighted ontology establishment algorithm was also designed to realize the dynamic, accurate, and comprehensive description of user interests. Through individual filtering, comprehensive filtering, concept mapping, semantic query expansion, and query result optimization, the recommendation and retrieval algorithm of the personalized product catalog information was constructed to establish the product catalog intelligence service system. In the contrasting experiment of the search results, compared with three used retrieval modes of the retrieval based on domain ontology, the optimized retrieval based on semantic similarity, and the optimized retrieval based on personalized ontology, the degree of the users＇ satisfaction increased by 31%, 21%, and 14% respectively. The experiment＇s result showed that the catalog intelligence service system could improve the quality and satisfaction of the pollution-free agricultural products directory information retrieval and recommendation to a certain extent, and met the personalized requirements. Meanwhile, the system also provided the technical support for the promotion, the brand sales, and electronic trading of the pollution-free agricultural product and showed a good prospect for application.