CAJ | 학술논문

美食资源库是个性化菜谱查询、营养推荐、疾病食疗的底层基础。针对国内目前还没有一个完善的中文美食开放连接资源库，构建了国内首个中文美食开放链接资源库并提供SPARQL查询和普通查询服务，为上层智能应用的开发提供底层平台。针对结构化数据较少的情况，对传统的TF-IDF算法进行改进，引入VIPS算法，提出针对半结构化美食网站的通用美食爬虫，使美食数据的抽取更加智能化，准确率提高22．1％。
미식자원고시개성화채보사순、영양추천、질병식료적저층기출。침대국내목전환몰유일개완선적중문미식개방련접자원고，구건료국내수개중문미식개방련접자원고병제공SPARQL사순화보통사순복무，위상층지능응용적개발제공저층평태。침대결구화수거교소적정황，대전통적TF-IDF산법진행개진，인입VIPS산법，제출침대반결구화미식망참적통용미식파충，사미식수거적추취경가지능화，준학솔제고22．1％。
Food linked open data is the basis of personalised menu inquiry, nutritional recommendations, disease diet and some other ap-plications.Since there is not a complete Chinese food linked open data yet, we established the first national Chinese food linked open data which offers SPARQL inquiry and general inquiry services.Moreover, it provided the underlying platform for the development of the upper in-telligent application.In light of the situation of lacking structured data, we also improved the traditional term frequency-inverse document fre-quency algorithm, introduced vision-based page segmentation algorithm, and proposed a general food crawler for semi-structured food website, this made the extraction of food data more intelligent, and improved the accuracy up to 22.1%.