建構以情感為基礎之線上信任評論模式

2014年1111人力銀行調查顯示，高達9成7之受訪者表示，在旅遊前會先上網參考其他網友撰寫之口碑、評價及相關訊息，由此可知，網路評價及口碑行銷對消費者具有一定之影響力。然而，目前線上評論內容過於分散，導致使用者需要花費許多時間閱讀每位評論者撰寫之內容，此外，推薦平台多以五個等級之評等方式為商品整體性評等，然後此種衡量模式缺少最重要之信任因子支持，此外，67%之消費者在購物前僅會閱讀六篇以下之評論。本研究試圖找出其他會影響使用者做決策時之關鍵因子，考量完整之影響因子來建構有效評論模式，然而，透過文獻探討瞭解評論內容之情感因素重要性，亦發現雖然信任對於網路評論之重要性，卻僅提出影響信任程度之因子。本研究基於正負面詞彙對評論之文字內容進行情感分析，並以展望理論之邏輯為基礎延伸至信任概念，假設平台使用者在對於閱讀評論時並無特定偏好(亦即皆為中立角度)，在無固定參照點(Reference Point)之情況下，若評論內容被信任程度較高，其風險感知程度將降低，反之，若評論內容被信任程度較低，其風險感知程度將提高，以納入量化信任之概念建構出本模式。由於推薦平台上之資訊過於龐大，針對資料分析，本研究於TripAdvisor之網路平台上，選取拉斯維加斯271間飯店中前十名飯店為研究對象，並選擇旅遊旺季2015年一至二月份為評論收集區間。透過資料分析，發現十間飯店經由情感分析調節後，整體評等皆呈現下調趨勢，但僅有Skylofts at MGM Grand下調幅度高達一分，然而，再納入信任因子調節後，發現有九名飯店之整體評等皆下調兩分以上，在飯店排序方面亦有大幅度之變動，如原先排序第六位之Four Seasons Hotel擠進冠軍寶座，更發現到原透過情感分析調節後排序第一之Staybridge Suites飯店，在經由本模式調節後，其排序卻僅居於第九位，換言之，相較於僅衡量情感因子，加入信任因子更加顯著地影響著飯店排名。此外，本研究基於評論內容所呈現之情感分為正面與負面評論，以及基於評論標章將評論分為可信度高與可信度低，發現本模式深受情感與信任相互影響，其中，負面情緒與可信度低之評論影響飯店排序更加顯著。過去關於口碑之研究以情感分析居多，在思考評論對於平台使用者之有用性時忽略了量化可信度之討論，因此，本研究期望融合情感分析與量化信任之兩大構面以建構出新的衡量模式，於實務上給予管理者更多參考依據與經營方向調整。

關鍵字

情感分析；網路評論模式；展望理論

並列摘要

A survey from 1111 Job Bank (Taiwan) in 2014 shows around 97% participants read online reviews and rankings regarding travel information. Online reviews and word-of-mouth are extremely important to consumers. However, online reviews are complex nowadays and users need to spend much time on reading from various sources. In addition, online platform like TripAdvisor only uses five levels for reviewers which lack credibility. 67％ of users read no more six reviews before consumption. As a result, this research attempts to discover key factors that affect customer decision making. We propose a model by using sentiment analysis in terms of positive and negative words and the concept of credibility inferred from Prospect Theory. We assume all users are neutral in reading the reviews (no fixed reference point). That is, higher reliable reviews may have lower risk, and vice versa. This research uses TripAdvisor to examine the proposed model and selects 10 out of 271 hotels in Las Vegas from Jan. to Feb. in 2015. This is also the peak season for traveling in Las Vegas. We discovered the overall ranking decreased of 10 hotels through sentiment analysis. However, only Skylofts at MGM Grand reduce 1 level. The rest of 9 hotels reduced 2 or more levels after considering the factor of credibility. Moreover, the factor of credibility affected overall ranking. The ranking of Four Season Hotel increased from 6 to 1. The ranking of Staybridge Suites Hotel decreased from 1 to 9. Apparently, the factor of credibility has greater influence on hotels’ ranking than sentiment analysis. This research discovers negative emotion and low credibility reviews have more influence on the hotels’ ranking. In this research, we combine sentiment analysis with credibility in the proposed model and provide more clues to enterprises on operation, management, and strategic decisions.

並列關鍵字

Sentiment Analysis ； Online Review Comments ； Prospect Theory

參考文獻

[1]Arndt, J. (1967). Role of product-related conversations in the diffusion of a new product. Journal of Marketing Research, 4, 291-295.

[2]Ba, S., & Pavlou, P. A. (2002). Evidence of the effect of trust building technology in electronic markets: Price premiums and buyer behavior. MIS Quarterly, 26(3), 243-268.

[4]Berger, J., & Schwartz, E. M. (2011). What drives immediate and ongoing word of mouth? Journal of Marketing Research, 48(5), 869-880.

[5]Bianchi, C., & Andrews, L. (2012). Risk, trust, and consumer online purchasing behaviour: A chilean perspective. International Marketing Review, 29(3), 253-275.

[6]Bickart, B., & Schindler, R. M. (2001). Internet forums as influential sources of consumer information. Journal of Interactive Marketing, 15(3), 31-40.

國際替代計量

建構以情感為基礎之線上信任評論模式

全文下載

主題瀏覽