發(fā)布時間：2018-05-20 04:38

  本文選題：倒排表 + 范圍查詢��； 參考：《燕山大學(xué)學(xué)報》2017年05期

【摘要】：針對星型組查詢只考慮了用戶的位置信息和朋友關(guān)系,而沒有考慮用戶的偏好信息,使得推薦質(zhì)量不高的問題,本文提出一種基于范圍的星型組查詢算法(m SG),同時考慮用戶的社交關(guān)系、文本信息和當(dāng)前位置信息,在給定范圍為用戶返回k個得分最高的星型組。為了迅速查找中心點用戶,提出帶倒排表的網(wǎng)格索引結(jié)構(gòu)。為中心點用戶更快速地查找朋友,提出了社交關(guān)系和文本信息混合索引結(jié)構(gòu)。最后,通過實驗表明,利用所提索引結(jié)構(gòu)減少了查詢算法的運行時間。
[Abstract]:The star group query only considers the user's location information and the friend relationship, but not the user's preference information, which makes the recommendation quality not high. In this paper, a range based star group query algorithm is proposed, which takes into account the user's social relations, text information and current location information, and returns k star groups with the highest score in a given range. In order to find central users quickly, a grid index structure with inverted table is proposed. A hybrid index structure of social relations and text information is proposed for center point users to find friends more quickly. Finally, the experimental results show that the proposed index structure can reduce the running time of the query algorithm.
【作者單位】： 燕山大學(xué)信息科學(xué)與工程學(xué)院;河北省計算機虛擬技術(shù)與系統(tǒng)集成重點實驗室;
【基金】：河北省自然科學(xué)基金資助項目(F2017203019)
【分類號】：C912.1;TP301.6
，

本文編號：1913243