運用關聯分析探勘民眾關注議題與發展方向：以環保議題為例

關聯分析近年來被應用於許多不同研究領域，在巨量資料中，探勘資料間之相互關係與規則。本研究目標在於運用關聯分析，從巨量非結構化資料中，探勘民眾關注之議題，並分析未來可能發展方向。我們以環保領域資料為研究樣本，建構環保議題偵測模型，探勘民眾目前關注之議題與未來可能討論風向。分析結果顯示，探勘PTT電子佈告欄之文章，可有效瞭解民眾關注議題，以及準確預測未來議題發展方向。透過此環保議題偵測系統，不但可以讓環保機關精準掌握輿情焦點，提高施政品質，也可強化政策宣導內容，大幅涵蓋與解答民眾所關注之議題。此分析系統也可以應用在不同領域，例如：產品市場調查與口碑分析。

Association analysis has attracted considerable attention recently in many research fields, mining data relations and rules from huge volume of data especially. This study aims at mining issues of public concern and analyzing its relations from massive of unstructured data. The main resource of this study is environmental related documents from PTT bulletin board system. A model is constructed via the collected environmental documents for predictions of issues of public concerns and possible future directions. The experimental results show that mining information from documents of PTT bulletin board system can effectively understand the public concerns and predict possible future directions. The reports from the prediction system may be used as a reference for environmental authorities. The prediction model we propose not only precisely masters of opinions from public to improve the administrative quality of environmental authorities, but also strengthens the content of press release to cover and answer the significant important issues of public concerns. The prediction system can be also applied to different applications, such as market investigation and opinion analysis.