- 學位論文
以模擬股市交易機制衡量部落格價值之初探研究
Experimental Study of Developing a Virtual Stock Market for Ranking Blogs
摘要
近年來因為網路普及度大幅提升，以及部落格風潮發展快速，部落格已經成為擁有強大影響力的平民媒體，並且有許多商業化成功的例子。部落格本身與一般網站不同的特性，衍伸出有別於一般網站的衡量、排名、與搜尋機制。另一方面，預測市場理論已經在商業、公司營運、政治、與運動等領域，展現了有效的輔助功能和準確的預測能力。本研究提出一個應用預測市場理論在建置部落格衡量與排名之虛擬股票市場，期能透過玩家之間的交易行為顯現部落格的相對價值，並利用視覺化的方式建構玩家之間的交易關係。 本研究建置了一個虛擬股票市場，分別以部落格和校園歌唱比賽進行了兩個實驗，主要目的在對該系統中的交易機制做測試、改良、與探討。另外為了指出使用多重帳號而獲利的玩家，本研究用視覺化的方式將可疑玩家的交易紀錄畫成交易關係網路圖，期能快速地找出可疑的交易行為。 本研究的結論有三點。第一，虛擬股票市場可應用在部落格推薦與排名上，但是需要時間與推廣讓使用者接受這個新概念。第二，實驗過程中發現，虛擬股票市場可以找出直覺上無法被數值化之物件在市場中的相對價值，例如歌手股價。第三，透過視覺化的交易關係網路圖，可以輔助找出在實驗中進行套利交易的作弊玩家。
並列摘要
Because of the mass growth of Internet popularity, and the quick spread of blogs in the recent years, blogs has become a civilian media with powerful influence. Several mechanisms have been developed for measuring, ranking, and searching blogs based on the different characteristics of blogs from original websites. On the other hand, the theory of prediction market has shown great abilities of predictions in business, organization managing, politics, and sports. This study proposes to implement a virtual stock market for valuing and ranking blogs. This study expects to reveal the relative value of listed blogs in the market by making transactions, and observes trading relationships among players by visualization tools. This study aims to implement a virtual stock market, and run two experiments with blogs and campus singing contest respectively, whose main purposes are testing, improving, and discussing the trading mechanism of the system. In order to identify the players who had profit by controlling multiple accounts, this study expects to find out suspicious trading behaviors by transforming target players’ transaction records into visualized trading relation networks. There are three conclusions of this study. First, it is workable to implement virtual stock market as a tool for blog recommendation and ranking, but it takes time to convince users to accept this new idea. Second, this study shows an extended ability of finding relative value of objects which is basically impossible to be quantified, such as the stock prices of singers. Third, this study shows that it is helpful to identify players with interest arbitrage in the experiment by drawing the trading relation networks.