【摘要】：隨著互聯(lián)網(wǎng)的飛速發(fā)展,社交網(wǎng)絡(luò)成為了覆蓋群體最廣、社會影響最大、商用價值最高的Web2.0服務(wù)。移動終端設(shè)備上定位技術(shù)的成熟與提升,極大地改善了基于位置的服務(wù)的質(zhì)量。社交網(wǎng)絡(luò)與基于位置的服務(wù)之間的融合誕生出了一種新型的社交網(wǎng)絡(luò)服務(wù)——基于位置的社交網(wǎng)絡(luò)�；谖恢玫纳缃痪W(wǎng)絡(luò)的顯著特點是能夠為用戶記錄帶有時間戳的位置信息。借助位置信息,基于位置的社交網(wǎng)絡(luò)的出現(xiàn)使虛擬世界與現(xiàn)實世界之間建立了更加緊密的關(guān)系。目前,基于位置的社交網(wǎng)絡(luò)已經(jīng)成為社交研究領(lǐng)域的一個新方向,它為相關(guān)的研究提供了大量的包含了時間、空間和社交關(guān)系的用戶位置簽到數(shù)據(jù)�；谖恢玫纳缃痪W(wǎng)絡(luò)的各類服務(wù),也為用戶行為預(yù)測和推薦系統(tǒng)等相關(guān)研究提供了更多的應(yīng)用點。本文的研究對象是基于位置的社交網(wǎng)絡(luò)中用戶的位置簽到數(shù)據(jù),以及相關(guān)的用戶注冊位置信息、線上社交關(guān)系網(wǎng)絡(luò)等數(shù)據(jù)內(nèi)容。通過對基于位置社交網(wǎng)絡(luò)中的用戶的線上行為特征的分析,發(fā)現(xiàn)了用戶位置簽到行為在時間、空間與社交關(guān)系上的多種特征,并依此建立用戶行為特征模型。在相應(yīng)的應(yīng)用場景中,本文提出了一套個性化位置推薦系統(tǒng),,給出了系統(tǒng)的詳細設(shè)計。該系統(tǒng)根據(jù)用戶特征模型,預(yù)測用戶潛在的興趣位置,為用戶推薦出行建議,提升了用戶在基于位置的社交網(wǎng)絡(luò)中享受位置推薦服務(wù)的體驗。本文基于用戶線上同行行為的定義,提出了面向同行群組的位置群推薦算法。在多人同行場景下,與面向個人的推薦算法相比,提升了推薦算法的準確度。該算法,融合了本文提出的用戶個性化位置推薦算法。測試了不同聚合方案中位置群推薦算法的性能,并基于分布式計算平臺Spark進行了系統(tǒng)實現(xiàn),提高了面向同行群組的位置群推薦性能。基于理論分析模型,本文提出了一個基于用戶位置行為特征的遠程提問系統(tǒng)。最終,針對本文的理論與實驗兩個方面給出了改進方向。
[Abstract]:With the rapid development of the Internet, social network has become the most widely covered group, the largest social impact, the highest commercial value of Web2.0 services. The maturity and improvement of location technology in mobile terminal devices greatly improve the quality of location-based services. The fusion of social networks and location-based services has given birth to a new type of social-networking service, location-based social network. A prominent feature of location-based social networks is the ability to record timestamp location information for users. With the help of location information, the emergence of location-based social networks makes the virtual world and the real world more closely related. At present, location-based social network has become a new direction in the field of social research. It provides a large number of user location check-in data including time, space and social relations for related research. The various services of location-based social networks also provide more application points for user behavior prediction and recommendation systems. The object of this paper is the location check in data of the user in the location-based social network, and the related information of the user registration location, the online social relationship network and other data content. By analyzing the online behavior characteristics of users in location-based social networks, this paper finds out a variety of features of user location-check-in behavior in time, space and social relations, and establishes a user behavior feature model. In the corresponding application scenario, this paper presents a personalized location recommendation system, and gives the detailed design of the system. According to the user characteristic model, the system predicts the potential location of interest of users, recommends travel advice for users, and improves the experience of users enjoying location recommendation services in location-based social networks. Based on the definition of peer behavior on user line, this paper proposes a location group recommendation algorithm for peer groups. Compared with the personal-oriented recommendation algorithm, the accuracy of the recommendation algorithm is improved in the multi-person peer scenario. This algorithm combines the user personalized location recommendation algorithm proposed in this paper. The performance of location group recommendation algorithm in different aggregation schemes is tested and implemented based on distributed computing platform Spark to improve the performance of position group recommendation for peer groups. Based on the theoretical analysis model, this paper presents a remote questioning system based on user location behavior characteristics. Finally, the direction of improvement is given in view of the theory and experiment of this paper.
【學位授予單位】：北京郵電大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：TP391.3

【參考文獻】

相關(guān)期刊論文 前1條

1 杜武恭;杜惠英;;個性化的周期性定位策略研究[J];互聯(lián)網(wǎng)天地;2015年07期

，

本文編號：2417208