本文選題：社會化標(biāo)注系統(tǒng)　切入點(diǎn)：標(biāo)簽聚類　出處：《情報(bào)雜志》2016年11期

【摘要】：[目的/意義]標(biāo)簽聚類能夠發(fā)現(xiàn)標(biāo)簽群體中蘊(yùn)含的知識和語義結(jié)構(gòu),從而緩解標(biāo)簽所面臨的歧義、模糊等問題,對于提升資源的檢索效率、改善用戶的使用體驗(yàn)、促進(jìn)社會化標(biāo)注系統(tǒng)的深化應(yīng)用具有十分重要的意義。[過程/方法]提出一種基于資源內(nèi)容聚類的標(biāo)簽聚類方法,該方法首先利用譜聚類算法對資源的詞特征進(jìn)行聚類,獲取資源內(nèi)容的K個(gè)特征簇,然后利用點(diǎn)互信息測量標(biāo)簽與這K特征簇的相關(guān)性,最后依據(jù)最大相關(guān)性原則將標(biāo)簽全體聚類成K個(gè)類簇。[結(jié)果/結(jié)論]實(shí)驗(yàn)結(jié)果表明,由于有效利用了資源的內(nèi)容這一重要信息,提出的方法與基于VSM的K-M eans聚類方法和基于VSM的凝聚式層次聚類方法相比,獲取了更好的聚類效果。
[Abstract]:[objective / meaning] tag clustering can find out the knowledge and semantic structure contained in tag group, so as to alleviate the ambiguity and ambiguity faced by tag, and improve the efficiency of resource retrieval and user's experience.It is of great significance to promote the application of social tagging system.[process / method] A label clustering method based on resource content clustering is proposed. Firstly, the spectral clustering algorithm is used to cluster the word features of resources, and K feature clusters of resource content are obtained.Then the correlation between the label and the K feature cluster is measured by using the point mutual information. Finally, all the tags are clustered into K clusters according to the principle of maximum correlation.[results / conclusion] the experimental results show that the proposed method is more effective than K-M eans clustering method based on VSM and condensed hierarchical clustering method based on VSM because of the effective use of the important information of resource content.
【作者單位】： 安徽理工大學(xué)經(jīng)濟(jì)與管理學(xué)院;復(fù)旦大學(xué)管理科學(xué)與工程博士后流動(dòng)站;
【基金】：國家自然科學(xué)基金項(xiàng)目“基于領(lǐng)域本體的煤礦安全數(shù)據(jù)融合方法及應(yīng)用”(編號:51474007) 教育部人文社會科學(xué)研究項(xiàng)目“社會化標(biāo)注環(huán)境下的標(biāo)簽層次關(guān)系發(fā)現(xiàn)方法研究”(編號:13YJCZH077) 安徽省人文社會科學(xué)重點(diǎn)研究基地項(xiàng)目“基于多源異構(gòu)信息融合的煤礦安全管理體系與方法研究”(編號:SK2014A042)和“煤礦安全風(fēng)險(xiǎn)領(lǐng)域本體的構(gòu)建技術(shù)及應(yīng)用研究”(編號:SK2015A082)資助
【分類號】：TP391.1
，

本文編號：1716338