因為網際網路的快速發展,廠商與消費者可以從網路上獲得具有參考價值的評論文章,但是,閱覽許多文章非常耗費時間,這也是本研究要改善的方向。 本研究建置一套中文意見探勘系統,針對特定領域進行分析,本論文運用演算法對論壇網站Mobile01上的電信領域文章進行分析,同時搭配人工針對演算法結果進行意見詞的標記工作,以提高意見詞的正確性,並提供分析圖表,使User可快速取得正確且客觀的資訊。 In the wake of rapid internet developing, review articles of reference value can be obtained from networks by both makers and consumers, however, the reading of the massive articles becomes very time consuming, and this is what the study aims at. The study established a Chinese opinion mining system, where algorithms were employed specific to special domain, analysis conducted, meanwhile matched with manual markup job for opinion words on algorithm results, to enhance accuracy of opinion words, and provide analysis diagrams, so that accurate and objective information can be obtained by users rapidly.