CAJ | 학술논문

微博作为最大的社会化媒体产品，拥有海量的用户和信息资源。微博推荐是微博个性化服务的重要方面，是解决信息过载问题的有效工具。考虑到微博数据海量性的特点，针对传统串行推荐算法对大数据处理效率低的问题，采用MapReduce模型，提出和设计一种基于关联规则挖掘算法Apriori的微博推荐并行算法，并在Hadoop平台实现。实验表明，提出的微博推荐并行算法具有较好的加速比和较高的运行效率，证明了该微博推荐并行算法在大数据处理中的高效性。
미박작위최대적사회화매체산품，옹유해량적용호화신식자원。미박추천시미박개성화복무적중요방면，시해결신식과재문제적유효공구。고필도미박수거해량성적특점，침대전통천행추천산법대대수거처리효솔저적문제，채용MapReduce모형，제출화설계일충기우관련규칙알굴산법Apriori적미박추천병행산법，병재Hadoop평태실현。실험표명，제출적미박추천병행산법구유교호적가속비화교고적운행효솔，증명료해미박추천병행산법재대수거처리중적고효성。
As the largest social media product, microblog has massive users and information resources.Microblog recommendation is an important aspect of personalized service, and it is an effective tool to solve the problem of information overload.Considering the massive microblog data and aiming at this problem of the traditional serial recommendation algorithms are inefficient when dealing with large data, a parallel algorithm of microblog recommendation based on Apriori which is an association rule mining algorithm is proposed and designed using MapReduce model, and it is achieved on the platform of Hadoop.The experiments show that the proposed parallel algorithm of microblog recommendation has the better speedup and the higher operating efficiency, proving the high efficiency of the parallel algorithm of microblog recommendation in dealing with large data.