CAJ | 학술논문

将社会化标签之间的关系抽象为一种复杂网络。将复杂网络中用于发现社区的连边社团检测算法应用于社会化标签的聚类分析中，利用其能够发现重叠簇的特性，以解决标签的语义模糊性，同时有效缓解了数据稀疏性问题。最后，以形成的标签簇为中介，通过计算用户对标签簇的兴趣度和资源与标签簇的相关度实现个性化资源推荐。实验结果表明了所提模型的优越性。
장사회화표첨지간적관계추상위일충복잡망락。장복잡망락중용우발현사구적련변사단검측산법응용우사회화표첨적취류분석중，이용기능구발현중첩족적특성，이해결표첨적어의모호성，동시유효완해료수거희소성문제。최후，이형성적표첨족위중개，통과계산용호대표첨족적흥취도화자원여표첨족적상관도실현개성화자원추천。실험결과표명료소제모형적우월성。
In this paper, the relationship between social tags was seen as a complex network. The algorithm of AL, which is used for finding communities in complex networks, was used to cluster social tags, the semantic ambiguity of tags was solved afterwards by making use of its specialty in finding overlapping clustering, at the same time, the problem of data sparsity was alleviated. Then, by computing user's interest for tag cluster formed and computing the correlation between resources and tag cluster, the paper achieved personalized rec-ommendation. The superiority of the model proposed was proved by experiment at last.