由於網際網路的快速發展,在網路上消費變得越來越方便,也越來越常見。現今消費者在網路消費前,會從社群網站的公開討論上比較商品間的差異,這已變成必做的功課。每個消費者對於不同的品牌都有著不同的看法。對於品牌方而言,一個好的行銷策略就會是一件非常重要的事。有鑒於此,本研究將應用文本探勘的技術對社群網站上的發言進行歸類,以便快速發掘消費者在品牌態度與購買行為上的相關因素。本研究首先將品牌行銷與市場行銷的問題歸納為構面(大標籤)和因素(小標籤);並利用這些標籤來建構文本探勘所需的專業詞庫。然後利用決策樹來建構文本多標籤分類模型,最後利用文字雲來做資料視覺化。本研究所提出的方法可讓品牌方更方便、快速地了解到品牌在消費者心中的模樣,對品牌未來的發展有重要的決定性影響。
Due to the rapid development of the Internet, online shopping has become increasingly convenient and common. Nowadays, before making an online purchase, consumers tend to compare the differences between products through discussions on social networking sites, which has become a necessary homework. Each consumer has a different opinion on different brands, and for the brand side, a good marketing strategy is crucial. For the reasons stated above, this study applies text mining techniques to classify user comments on social networking sites, aiming to quickly discover the relevant factors in consumer attitudes and purchasing behavior towards brands. This study begins by categorizing the issues of brand marketing and market marketing into dimensions (big labels) and factors (small labels). These labels are then used to construct a specialized lexicon required for text mining. Subsequently, a decision tree is utilized to build a multi-label classification model for the text. Finally, a word cloud is used for data visualization. The proposed method in this study enables brand owners to understand the perception of their brand conveniently and quickly in consumers' minds, which has a decisive impact on the future development of the brand.