2017 年,軍公教年金改革議題炒得沸騰,各家媒體對此事件做了許多不 同的報導與回應。本研究欲了解媒體的框架效果,因此特別針對台灣新聞媒 體《自由時報》、《蘋果日報》、《聯合報》,以及 PTT 社群,透過語意 網絡分析、字詞情感分析等研究方法 , 期待掌握新聞對民眾的網路輿情趨勢。 研究發現,媒體的報導與談論方向確實受自身立場有所差異。此外,透過語 意網絡圖和字詞情緒分析,亦將媒體的動態框架化過程清楚呈現。

關鍵詞：年金改革、框架效果、語意網絡分析、輿情分析、社群媒體、新聞媒體

In 2017, the issue of pension reform was so noisy, and news media and people were kept prevailing different views about it. In order to realize the farming effect of mediums, I chose Taiwan's mainstream media, ''Liberty Times', "United Daily New", "Apple Daily', and "PPT" as my research targets. Using semantic network analysis, Text sentiment analysis and path analysis, expect to have a deep understanding about framing theory via monitoring what media tried to say and how the audience responded about the issue. The research found that each media has constrained perspective, so they would discuss the pension reform issue according to their standpoints, which verified the framing effect. Furthermore, semantic network graph visualized popular topics separately, sentiment and regression analysis also helped realize the dynamic process of media framing.

Keywords: Pension Reform, Semantic Network Analysis, Opinion Mining, Framing Effect, News Media, Social Media