在社群網路服務中找出不活躍的使用者 ; Discovering Inactive Users in Social Network Service

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/44729

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/44729

Title:	在社群網路服務中找出不活躍的使用者;Discovering Inactive Users in Social Network Service
Authors:	徐誌良;Chih-Liang Hsu
Contributors:	資訊工程研究所
Keywords:	時間的距離;離群偵測;社群網路服務;Outlier Detection;Social Network Service;Time Distance
Date:	2010-07-28
Issue Date:	2010-12-09 13:53:54 (UTC+8)
Publisher:	國立中央大學
Abstract:	社群網路服務在近年來變成熱門的服務, 有越來越多類似的服務在網際網路上開始運作。由於社群網路服務種類非常的多, 使得新手使用者不知道該如何去選擇適合自己的服務。許多使用者沒有從中得到樂趣或是根本不知道使用這個服務的好處在哪。舉個兩個例子來說, 第一種是剛加入社群網路服務的使用者, 他基本上不會知道有誰適合做自己的朋友, 更不會知道有哪些社團是自己可能會感興趣的, 當然跟別人的互動就少。第二種是有些使用者在加入了服務一段時間之後, 因為找不到志同道合的朋友或是找不到這個服務的樂趣在哪, 最後放棄了使用這個社群服務。這兩種使用者可以當作社群網路中的弱勢團體, 通常傾向不活躍的使用者, 因為他們跟其他使用者的互動比較少, 進而導致慢慢的放棄這個服務。為了改善這種現象, 我們期許能夠幫助這些使用者去發掘出社群網路服務的吸引力。一般來說, 一個剛踏入某個社群網路服務的使用者通常充滿著無限的潛力, 因為他們正在熱衷的階段, 也最希望能在這個服務中嘗試新鮮的事物。假如這個新的使用者在使用初期受到了好的指引, 他可能從此愛上這個服務。因此這篇論文對分析社群網路的資料, 提出了一個新的思維。為了要幫助首先我們先找出每個使用者之間發文和回文的時間差異的值, 作為時間的距離, 把這些數值去做處理之後, 去產生出使用者之間的距離。有了這些使用者之間的距離之後, 我們可以更近一步的去做以距離為基礎的離群偵測演算法。這個演算法可以利用點與點之間距離, 去尋找出在某個社群中的離群者。而這些社群中的離群者, 就是我們之前提到的不活躍的使用者。我們可以針對這些不活躍的使用者, 去觀察他們在這個社群服務中的喜好, 像是傾向哪種的分享資源或是偏愛哪一個使用者的發文。有了這方面的資訊之後, 對於該如何提升使用者使用社群網路服務的意願, 會有很大的幫助。Social network service (SNS) has become the hot service in recent years. More and more similar services are created on the website. The kinds of services are too lots that many users don’t know how to choose, and some users in the SNS don’t find the enjoyment even after they have used the service. For example, the new members don’t know which guys he/she would like to meet or which groups he/she would feel interesting. In order to improve this phenomenon, we decide to help these users to unearth the attraction of the SNS. Another point is that the major revenue of the social network service is the advertisement on the website. Therefore, it’s significant to think over the effective method to increase the watched times of the advertisement for the designers of the services. However, suppose the entire users see the same kinds of advertisement on their browser, it wouldn’t reach the effective propagation. In other words, our goal is to let different users could see the appropriate advertisement which they may interest on the website. Hence, this paper designs a novel method to analyze the data of SNS on website. At first, we find the time distance between the user nodes, and use these values to represent the distance between the users. In the second, in order to help the users to unearth the attraction of the SNS, we must to find out the target users in the SNS on website first, and then we use these data of distance between the users to run the distance-based algorithm of outlier detection. We called the result data after we computing as inactive users, we can analyze the information of the resources to find which resources do the inactive users interesting and the users who sharing the resources to the inactive users in the SNS on website. The result of the analysis can help us to recommend to the users about which members or groups are better to meet and join.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	599	View/Open

社群 sharing

Loading...