網路訊息的關鍵字與情感分析─以長榮罷工事件為例

隨著社群網路科技的進步，民眾經常在網路上表達意見。這些網路訊息可在新聞網站、討論區文章或社群媒體加以擷取，並以文本探勘或自然語言處理技術做關鍵字及情感分析。分析結果有助於瞭解民眾對社會事件的看法，發掘其隱含意義。本研究以2019年6、7月間發生的長榮航空空服員罷工事件為例，從 PTT八卦板與航空板爬取文本，進行網路訊息的關鍵字分析與情感分析，以探討此一事件之相關社會議題。研究結果顯示，長榮航空空服員罷工事件主要訴求以保障勞方權益為主，希望強化工會功能，參與人事決策，並提出參加工會者方得享有權益的禁搭便車條款。在社會觀感方面，民眾的負面情感強於正面情感，但總聲量卻日益減弱。經過整理並分類關鍵字之後，可將其分為勞方權益、企業權益、顧客權益、工會自主四大類別，其中以企業權益與顧客權益中的關鍵字較常出現，代表大眾在關注的議題上較為重視和自身相關的事件，並且立場上偏向支持資方。

關鍵字

文字探勘；情感分析；網路爬蟲；勞工權益；勞資糾紛

並列摘要

With the advancement of social networking technology, people often express their opinions online. These messages can be captured on news websites, forums or social media by using software technology, and their meaning explored through using text mining, keyword analysis, natural language processing or sentiment analysis. These analysis results can help us understand people's perceptions about public affairs. This study takes the EVA Air flight attendants' strike in June and July 2019 as an example, trying to analyze the public's perceptions of the strike from the posts of PTT. The results of the study show that the Evergreen Airline flight attendants' strike mainly focuses on protecting labor rights, strengthening power of the union. In terms of impressions from society, public sentiment has been more negative than positive, and the amount of posts has been decreasing over time. The keywords can be divided into four categories: labor rights, enterprise rights, customer rights, and union autonomy. Among them, keywords in enterprise rights and customer rights appear more often, representing the public pay more attention to the events related to themselves, and prefer to agree EVA’s claim about this strike.