線上評論是常見且重要的消費者回饋，會影響消費者的購買意圖，對廠商而言，線上評論可以幫助廠商找出產品或服務的優缺點，甚至能與其他品牌進行比較，以做為未來策略參考因素之一。然而現今線上評論繁雜，消費者與商家都無法在短時間內從線上評論中得到全面性的見解，故本研究以化妝品為例，設計一套「自動化情緒分析系統」，透過台灣大學NLPLab所建置的中文意見詞典ANTUSD更準確呈現使用者的情感強烈程度，並且幫助廠商快速找出消費者在意的產品屬性，以客觀的角度發現自身廠牌的優缺點，找出與其他廠牌相比的優劣勢，使廠商能夠迅速的回應市場需求，消費者也能透過本系統找出與自身膚質相符的產品。;Online reviews are common and important consumer feedback that will affect consumers′ intentions of buying. For manufacturers, online reviews can help manufacturers find the advantages and disadvantages of a product or service and even compare it with other brands as a future Strategy reference factor. However, today′s online reviews are large and complicated, consumers and businesses can not be in a short time from the online comments to get a comprehensive view, so this research about cosmetics, for example, the design of an "automated sentiment analysis system". It helps customers and manufacturers quickly to find out the aspects of consumers care about the product, giving an objective point of view to find the advantages and disadvantages of their own brand. The goal is to find out the advantages and disadvantages compared with other brands, so that manufacturers can quickly respond to market demand, consumers can find ideal product.
