Product characteristic weighting for designer from online reviews: an ordinal classification approach

Online product reviews are a reliable source of customers' sentiments. Directly connecting with customers and designers, online reviews can shorten product development life cycles. The problem arising is, although different techniques for processing online reviews are developed, the techniques are rarely seen on accelerating the design work flows. This paper proposes a two stage framework to learn the importance of characteristics from online reviews which could benefit product design. The first stage is a supervised learning routine to identify product characteristics. This procedure calculates the surrounding words' posterior probability. A linear weight learning algorithm is subsequently shown to reach the product characteristics identification. The second stage focus on meeting customers' needs. Distinct from existing classification and rank algorithms, this stage informs an ordinal classification algorithm to balance the weight of product characteristics. This algorithm depicts a pairwise approach to achieve ordinal classification. Furthermore, an integer none linear programming model is advised, which targets at recovering pairwise results to the original class for each object. Four brands of printer reviews from Amazon and real analysis from two experienced product designers are employed in this experimental study. The results validate the feasibility of the two stage framework, and the possibility to explore targeted models from online reviews for product designers.

[1]  Thorsten Joachims,et al.  Training linear SVMs in linear time , 2006, KDD '06.

[2]  Hong Yu,et al.  Towards Answering Opinion Questions: Separating Facts from Opinions and Identifying the Polarity of Opinion Sentences , 2003, EMNLP.

[3]  Tie-Yan Liu,et al.  Learning to rank for information retrieval , 2009, SIGIR.

[4]  Oren Etzioni,et al.  Extracting Product Features and Opinions from Reviews , 2005, HLT.

[5]  Tao Qin,et al.  Query-level loss functions for information retrieval , 2008, Inf. Process. Manag..

[6]  Gregory N. Hullender,et al.  Learning to rank using gradient descent , 2005, ICML.

[7]  Paolo Rosso,et al.  Overview of the 2nd international workshop on search and mining user-generated contents , 2010, CIKM '10.

[8]  Ying Liu,et al.  How to interpret the helpfulness of online product reviews: bridging the needs between customers and designers , 2010, SMUC '10.

[9]  Bing Liu,et al.  Mining Opinion Features in Customer Reviews , 2004, AAAI.

[10]  Ramesh Nallapati,et al.  Discriminative models for information retrieval , 2004, SIGIR '04.

[11]  Soo-Min Kim,et al.  Determining the Sentiment of Opinions , 2004, COLING.

[12]  Thomas Y. Lee,et al.  Needs-based analysis of online customer reviews , 2007, ICEC.

[13]  Tie-Yan Liu,et al.  Learning to Rank for Information Retrieval , 2011 .

[14]  Nigel Collier,et al.  Sentiment Analysis using Support Vector Machines with Diverse Information Sources , 2004, EMNLP.

[15]  Han Tong Loh,et al.  Gather customer concerns from online product reviews - A text summarization approach , 2009, Expert Syst. Appl..

[16]  Bei Yu,et al.  TSA'09 workshop summary: topic-sentiment analysis , 2009, CIKM.

[17]  Richard A. Harshman,et al.  Indexing by Latent Semantic Analysis , 1990, J. Am. Soc. Inf. Sci..

[18]  David M. Blei,et al.  Supervised Topic Models , 2007, NIPS.

[19]  Michael I. Jordan,et al.  Latent Dirichlet Allocation , 2001, J. Mach. Learn. Res..

[20]  Bing Liu,et al.  Mining and summarizing customer reviews , 2004, KDD.

[21]  Thorsten Joachims,et al.  Optimizing search engines using clickthrough data , 2002, KDD.

[22]  Lei Zhang,et al.  Entity discovery and assignment for opinion mining applications , 2009, KDD.

[23]  Michael L. Littman,et al.  Measuring praise and criticism: Inference of semantic orientation from association , 2003, TOIS.

[24]  Yoram Singer,et al.  An Efficient Boosting Algorithm for Combining Preferences by , 2013 .