CAJ | 학술논문

微博社区发现在舆情分析、个性化推荐等方面具有重要的应用价值。为了准确而高效地发现微博社交网络中的社区，提出了一种基于连边层次聚类的微博社区发现方法。该方法通过高度重叠社区的合并及划分误差的修正，进一步提高了微博社区发现的准确率。为了提高微博社区发现的效率，利用开源云计算平台Hadoop所提供的MapReduce编程模型进行了分布式并行处理。实验结果表明，所采用的微博社区发现方法不仅具有较高的准确率，而且具有较高的效率。
미박사구발현재여정분석、개성화추천등방면구유중요적응용개치。위료준학이고효지발현미박사교망락중적사구，제출료일충기우련변층차취류적미박사구발현방법。해방법통과고도중첩사구적합병급화분오차적수정，진일보제고료미박사구발현적준학솔。위료제고미박사구발현적효솔，이용개원운계산평태Hadoop소제공적MapReduce편정모형진행료분포식병행처리。실험결과표명，소채용적미박사구발현방법불부구유교고적준학솔，이차구유교고적효솔。
Micro-blog community detection has an important application value in public opinion analysis and person-alized recommendation, etc. In order to find communities from micro-blog social network accurately and efficiently, this paper proposes a method of micro-blog community detection based on hierarchical clustering of edge. This method further improves the accuracy of community detection through merging highly overlapping communities and correcting divided error. To improve the efficiency of micro-blog community detection, this paper carries out distributed parallel processing with MapReduce model provided by open source cloud computing platform—Hadoop. The experimental results show that the micro-blog community detection method has higher accuracy and efficiency.