عنوان انگلیسی مقاله:
Investigating consumer preferences on product designs by analyzing opinions from social networks using evidential reasoning
ترجمه فارسی عنوان مقاله:
بررسی ترجیحات مصرف کننده در طرح های محصول با تجزیه و تحلیل نظرات شبکه های اجتماعی با استفاده از استدلال مشهود
Sciencedirect - Elsevier - Computers & Industrial Engineering, 139 (2020) 106180: doi:10:1016/j:cie:2019:106180
C.Y. Nga,⁎, Kris M.Y. Lawb
The rapid growth of e-commerce and social networking sites has created various challenges for the extraction of
user-generated content (UGC). In the era of big data, customer opinions from social media are utilized for
investigating consumer preferences to support product redesigns.
Opinion mining, including the various automatic text classification algorithms using sentiment analysis is a
capable tool to deal with a large amount of comments on the social networking sites. In which, sentiment
analysis is used to determine the contextual polarity within a comment by searching sentimental words.
However, the inconsistency on choosing the sentiment words leads to the inaccurate interpretation of the opinion
strength of sentiment words.
An approach to summarize the UGC from social networking media using fuzzy and ER without the need to
review all the comments is proposed in this paper. The inaccuracy on determination of the polarity of sentiment
words and corresponding opinion strengths is rectified by fuzzy approximation and ER. The result is presented in
ranking therefore the effort for result interpretation significantly reduced.
The incorporation of sentiment analysis with ER to analyze the UGC for product designs is a new attempt in
investigating consumer preferences. The proposed approach is shown to be handy, sufficient, and cost effective
for the product design and re-design, particularly in the preliminary stage.
This project can be further extended by employing alternative fuzzy approximate techniques in the fuzzy-ER
approach to support the sentiment analysis to enhance the accuracy of sentiment values for determining the
distribution assessments of ER.
Keywords: Opinion mining | Sentiment analysis | Evidential reasoning | Consumer preferences | Product design