運用文字探勘技術於企業聲譽分析之研究-以企業社會責任為例

為持續保持競爭優勢，企業必需擁有良好的企業聲譽，而網路的發達及社群媒體的發展，讓許多企業高層開始重視社群媒體的影響力，希望能瞭解其在網路上的聲譽及形象。此外，近年來大眾相當看重企業對於整體經濟、公益活動、社會參與以及環境永續等社會責任相關議題，使得「企業社會責任」構面逐漸成為企業建立聲譽及形象的重要指標。因此，本研究以企業社會責任為例，透過文字探勘、深度學習技術為企業提出一適合衡量企業社會責任的指標及分析模組，藉由蒐集大眾於社群平台（PTT、FB）與企業相關新聞文本及評論留言進行新聞文本分類及情緒分析，來了解企業在網路的企業社會責任形象與聲譽。　　本研究針對網路搜集之企業新聞文本以「企業社會責任」構面進行標記，並使用SVM、CNN及LSTM三種方法進行分類、比較，找出構面分類成效較佳之方法；此外，針對各構面透過CNN、LSTM及Bi-LSTM方法將企業評論留言進行情緒分類，分別計算出企業在各構面評論留言之正負面情緒及情緒分數，以呈現企業在企業社會責任之表現與形象。　　經本研究搜集之企業「企業社會責任」資料與提出模型分析結果，研究結論主要如下：(1)在新聞文本分類模型中，SVM分類效果相對最為穩定；(2)在情緒分類模型中，Bi-LSTM分類成效最佳，準確率皆達80%以上；(3)透過情緒分析過程，亦發現不同資料來源會呈現出不同評論留言情緒傾向，PTT資料集的評論留言所呈現之情緒以負面居多、Facebook資料集的評論留言情緒表現則正負情緒涵蓋較為平衡；(4)將情緒分析各企業的情緒得分結果與天下雜誌「2019 天下CSR企業公民獎」之得獎排名進行比對後，證實本研究提出之模型架構可以有效用於分析企業在企業社會責任之表現，清楚了解網路大眾對企業在企業社會責任各構面之看法及情緒傾向，並藉此預測出企業間的社會責任排名及評定企業聲譽。

關鍵字

企業聲譽；企業社會責任；文字探勘；深度學習；情緒分析

並列摘要

In order to maintain the competitive advantage, the corporate must have a good corporate reputation. The development of the Internet and social media have led many corporate executives pay more attention to the influence of social media, hope to understand their reputation and image on the Internet. In recent years, the public has paid great attention to corporate social responsibility issues, such as overall economy, charitable event, social participation, and environmental sustainability, making the "Corporate Social Responsibility" dimensions become an important indicator of corporate reputation and image. Thus, this study takes Corporate Social Responsibility (CSR) as an example, through text mining and deep learning technology, proposes an index and analysis module for measuring CSR, collect news text and comment messages about corporations on the social media (PTT、FB), classify the news text, do sentiment analysis to understand the CSR image and reputation of the corporates on the Internet. 　　This study marks the news texts according to the CSR dimensions, and use SVM, CNN and LSTM three classification methods to find out the better classification method. At last, use CNN, LSTM and Bi-LSTM to sentiment classify the comment about the corporation, calculate sentiment scores on all the dimensions to show the CSR performance and image. 　　The analysis results from corporate CSR data and models through this study show that: (1) The SVM classification is relatively stable in the news text classification model. (2) The Bi-LSTM classification is the best in the sentiment classification model, and the accuracy rate is all more than 80%. (3) Through the process of sentiment analysis, it is also found that different data sources will show different sentiment tendencies. The sentiment of the PTT data set is mostly negative. In other hand, the sentiment of the FB data set is more balanced between positive and negative sentiments. (4) After comparing the sentiment score results of each company with the ranking of "Excellence in Corporate Social Responsibility Award" through the CommonWealth Magazine in 2019, it is confirmed that the model proposed in this study can effectively use to analyze the corporate performance of CSR. It can provide corporations as a reference to certainly understand the public's views and emotional tendencies on all the dimensions of CSR, also predict rankings between corporations and access corporation reputation.

並列關鍵字

Corporate Reputation ； Corporate Social Responsibility ； Text Mining ； Deep Learning ； Sentiment Analysis

參考文獻

Fombrun, C., Gardberg, N. A., Sever, J. M. (2000). The Reputation Quotient: A Multi-stakeholder Measure of Corporate Reputation. Journal of Brand Management, 7(4), 241-255. doi:10.1057/bm.2000.10

Gers, F. A., Schmidhuber, J., & Cummins, F. (1999). Learning to forget: continual prediction with LSTM. Neural Computation, 12(10), 2451-2471. doi: 10.1162/089976600300015015

Meijer, M. M., & Kleinnijenhuis, J. (2006). News and corporate reputation: Empirical findings from the Netherlands. Public Relations Review, 32(4), 341-348. doi: 10.1016/j.pubrev.2006.08.002

Roberts, P. W., & Dowling, G. R. (2002). Corporate reputation and sustained superior financial performance. Strategic Management Journal, 23(12), 1077-1093. doi: 10.1002/smj.274

Saxton, M. K. (1998). Where do Reputations Come From?. Corporate Reputation Review, 1(4), 393-399. doi: 10.1057/palgrave.crr.1540060

國際替代計量

運用文字探勘技術於企業聲譽分析之研究-以企業社會責任為例

全文下載

主題瀏覽