CAJ | 학술논문

传统的M ashup服务推荐是基于关键字的检索方法，对于所推荐 A PI服务的社会和功能属性利用较少，不利于全面评价所推荐A PI的适用度，对此提出一种多属性融合评级的M ashup服务推荐算法。首先，利用网爬工具收集ProgrammableWeb上的Mashup服务信息，并采用后缀剥离算法把Mashup服务的标签信息修改为名词形式，以此作为研究分析的数据集。其次，融入A PI服务的社会和功能等多属性对A PI模型进行扩充，并采用多属性相似度加权融合的方式对候选A PI的适用度进行评价，以此作为A PI服务推荐的依据。实验结果表明，多属性融合评级M ashup服务推荐算法具有更高的正确率和更快的运算时间，是可行有效的。
전통적M ashup복무추천시기우관건자적검색방법，대우소추천 A PI복무적사회화공능속성이용교소，불리우전면평개소추천A PI적괄용도，대차제출일충다속성융합평급적M ashup복무추천산법。수선，이용망파공구수집ProgrammableWeb상적Mashup복무신식，병채용후철박리산법파Mashup복무적표첨신식수개위명사형식，이차작위연구분석적수거집。기차，융입A PI복무적사회화공능등다속성대A PI모형진행확충，병채용다속성상사도가권융합적방식대후선A PI적괄용도진행평개，이차작위A PI복무추천적의거。실험결과표명，다속성융합평급M ashup복무추천산법구유경고적정학솔화경쾌적운산시간，시가행유효적。
The traditional Mashup service recommendation is a retrieval method based on keywords ,the use of social and functional properties for the recommended API service is too less to make a comprehensive e-valuation of the recommended API applicability ,so the Mashup service recommendation based on multi at-tributes fusion rating algorithm has been proposed to solve this problem .Firstly ,the climbing tools have been used to collect ProgrammableWeb Mashup service information ,and the suffix stripping algorithm been used to modify the Mashup service label with noun form as the research and analysis data sets .Sec-ondly ,the API model has been extended with the social and the functions attributes ,then the multiple at-tribute similarity weighted fusion has been used to evaluate candidate API fitness ,which would be the API service recommendation basis .The experimental results show that ,the multiple attribute fusion rating Mashup service recommendation algorithm has higher accuracy and faster computing time ,which is feasi-ble and effective .