本研究以台灣的政治議題作為研究主題，應用文本分析進行輿論聲量分析。本研究主要目的為透過文本分析，進行輿論聲量分析，以探討輿論聲量是否可以提前反映事實結果、了解各社群媒體是否有不同主觀立場。本研究首先利用網路爬蟲技術，擷取政治議題相關的網路新聞，接著進行斷字斷詞處理、TF-IDF權重計算，建立正、負輿論預測模式。研究結果發現，輿論聲量可以提前反應選舉的結果、不同社群媒體有各自特定立場，結合正、負輿論分析，可以預測某事件的正、負聲量。

This study takes Taiwan's political issues as the research theme and uses text analysis to analyze the volume of public opinion. The main purpose of this study is to analyze the volume of public opinion through text analysis, in order to explore whether the volume of public opinion can reflect the factual results in advance, and to understand whether various social media have different subjective positions. This study firstly uses web crawler technology to capture online news related to political issues, and then performs CKIP processing, TF-IDF weight calculation, and establishes a positive and negative public opinion prediction model. The results of the study found that the volume of public opinion can reflect the results of the election in advance, and different social media have their own specific positions. Combined with the analysis of positive and negative public opinion, the positive and negative volume of an event can be predicted.