摘 要:隨著社交網(wǎng)絡(luò)的發(fā)展,特別是伴隨著微博客這一新型社交網(wǎng)絡(luò)媒體的興起,社交平臺(tái)已經(jīng)成為用戶發(fā)布、獲取消息的最重要渠道之一。截止目前,僅新浪微博的注冊用戶數(shù)已經(jīng)超過5億人,日均信息發(fā)布數(shù)量已經(jīng)超過1億條。用戶可以發(fā)布不超過140字的博文,他們發(fā)布的內(nèi)容會(huì)出現(xiàn)在關(guān)注他們的用戶的時(shí)間線里。借助轉(zhuǎn)發(fā)功能,用戶可以使信息在用戶間產(chǎn)生滾雪球式級(jí)聯(lián)轉(zhuǎn)發(fā),從而令普通用戶也有可能產(chǎn)生巨大的影響力。通過將合適的用戶提及(@)在微博中,他們將會(huì)收到系統(tǒng)發(fā)出的微博提及提醒,他們可能的轉(zhuǎn)發(fā)將會(huì)幫助微博提升傳播力。該工作設(shè)計(jì)了一種全新算法,通過尋找最合適的提及對(duì)象來提升一條微博的傳播力。在研究過程中,我們深入調(diào)研了微博的提及機(jī)制,并且提出了一個(gè)推薦算法,來解決在微博中提及誰能使微博傳播力最大化的問題。在該工作中我們將微博提及的推薦問題,轉(zhuǎn)換為一個(gè)排序問題。該問題與傳統(tǒng)問題相比存在著四大全新挑戰(zhàn),包括:排序模型的相關(guān)性需要與信息傳播相關(guān);需要構(gòu)建基于話題的用戶關(guān)系模型;推薦存在嚴(yán)格的長度限制;推薦結(jié)果容易造成提及過載等問題。因此我們構(gòu)建了一個(gè)全新的排序模型:我們考慮了用戶興趣與微博內(nèi)容契合度、基于話題的用戶關(guān)系、以及用戶影響力指標(biāo)等三大類因素作為排序的特征;我們構(gòu)建了以信息傳播力為標(biāo)準(zhǔn)的新排序相關(guān)性模型;我們基于機(jī)器學(xué)習(xí)的方法,訓(xùn)練一個(gè)全新的排序函數(shù)。在實(shí)驗(yàn)過程中,我們搜集了來自新浪微博的大量真實(shí)的用戶信息,我們設(shè)計(jì)了多種對(duì)照算法,橫向測試了算法的表現(xiàn)。同時(shí),我們還針對(duì)算法使用的不同屬性的效用,針對(duì)推薦長度限制、推薦過載等問題,分別設(shè)計(jì)了對(duì)應(yīng)的實(shí)驗(yàn)。經(jīng)過詳盡的實(shí)驗(yàn)比較,我們提出的算法的表現(xiàn)要遠(yuǎn)優(yōu)于其他對(duì)照算法。我們的算法在只推薦極少量用戶的情況下,也能取得良好的推薦效果。同時(shí)我們設(shè)計(jì)的算法推薦結(jié)果分布較為平滑,不易出現(xiàn)推薦過載的問題。
關(guān)鍵詞:信息傳播 社交網(wǎng)絡(luò) 提及推薦 排序
Abstract:Nowadays, micro-blogging systems like Twitter have become one of the most important ways for information sharing. In Twitter, a user posts a message (tweet) and the others can forward the message (retweet). Mention is a new feature in micro-blogging systems. By mentioning users in a tweet, they will receive notifications and their possible retweets may help to initiate large cascade diffusion of the tweet. To enhance a tweet’s diffusion by finding the right persons to mention, we propose in this paper a novel recommendation scheme named as whom-to-mention. Specifically, we present an in-depth study of mention mechanism and propose a recommendation scheme to solve the essential question of whom to mention in a tweet. In this paper, whom-to-mention is formulated as a ranking problem and we try to address several new challenges which are not well studied in the traditional information retrieval tasks. By adopting features including user interest match, content-dependent user relationship and user influence, a machine learned ranking function is trained based on newly defined information diffusion based relevance. The extensive evaluation using data gathered from real users demonstrates the advantage of our proposed algorithm compared with the traditional recommendation methods.
Key Words:Information diffusion; Social network; Mention recommendation; Ranking
閱讀全文鏈接(需實(shí)名注冊):http://www.nstrs.cn/xiangxiBG.aspx?id=48557flag=1