主題事件趨勢研究

Abstract

主題事件趨勢長久以來都是各國政府及企業所關心的議題，透過主題事件趨勢預測，可以分析一個國家目前的社會經濟狀況及未來國家發展的走向，進而制定適當的決策。隨著資訊科技快速的演進，搜尋引擎已成為群眾在網路上蒐集資訊的重要工具，使用者於搜尋引擎上的行為模式與當前主題事件的發展息息相關，存在於搜尋引擎中的群眾智慧對於主題事件趨勢預測也深具研究發展的潛力。 在本研究中，我們基於隱含於搜尋引擎中的群眾智慧提出了協同式主題事件預測架構以評估主題事件當前的發展狀態。此架構包含兩個主要的貢獻，首先，我們提出了基於群眾智慧的特徵選取方法以找出具有代表性的特徵詞彙當作預測主題事件發展的關鍵指標。接著，我們進一步的基於搜尋引擎中的群眾智慧設計了一套新式的主題事件預測方法，以評估當前主題事件的發展狀態。實驗結果顯示，我們所提出的協同式主題事件預測架構能夠精準的進行預測，並且證明搜尋引擎中的群眾智慧能夠有效的運用於主題事件趨勢預測領域。

In response to rapidly changing situations at the national and international levels, it is important for decision makers to monitor the development of trending topics which are associated with long-running events that affect people’s lives and activities. In recent years, web search engines have become a major platform for the general public to access information. Because the search patterns of search engine users are often correlated with emerging events, the crowdsourcing of search engines has the potential for trend surveillance. In this dissertation, we provide a collaborative trend surveillance framework to estimate the status of trending topics by crowdsourcing the collective wisdom in web search engines. First, we propose a crowdsourced-wisdom-based feature selection method to select representative indicators showing trending topics and concerns of the general public. Then, we describe the design of our novel prediction method to estimate the trending topic statuses by crowdsourcing public opinion in web search engines. The experiment results show that the collaborative trend surveillance framework performs well and the crowdsourcing of web search engines are helpful for trend surveillance.