Exploring the Relationship between Customer Reviews and …lg5bt/files/Final Presentation...
Transcript of Exploring the Relationship between Customer Reviews and …lg5bt/files/Final Presentation...
Exploring the Relationship between
Customer Reviews and Prices
Lingjie Zhang, Lin Gong, Bo Man
Roadmap
● Introduction…………………………...Lingjie
● Methodology………………………….Lin
● Experimental Results…...…………...Bo
Customer Reviews Play an Important Role
90% customers say buying decisions are
influenced by online reviews.
Use of Customer Reviews
For customers
● Decision
● Recommendation
For retailers
● Feedback
● Marketing strategies
To what extend do they care about those reviews?
Motivation
Do customer reviews indirectly affect
sale prices?
Related Work
Classify reviews to help make decisions.
Extract opinion features in customer reviews.
Recommend products for customers.
None of them combine customer reviews with prices.
Challenge
● Relationship(Reviews,Prices)?
● Rating = Content?
Methodology
Step 1: Collect Reviews
SNAP Amazon reviews:
• Products with over 100 reviews, in total 419 products.
• Time period: Aug, 2012 - Mar, 2013
Step 2: Assumption
User ratings == User reviews
Machine Learning Methods are adopted.
(Naive Bayes, Logistics Regression, Support Vector Machine)
Given contents -> predict ratings.
Compare final precisions and recalls.
Prediction Results:
Naive Bayes
Step 3: Crawl Prices
Price data:
• 221 items from previous 419 items
• Time period: Oct, 2012 - Mar, 2013
Scaling:
Moving average:
Shift Analysis: • Compare against the prices ending L days
later than the ratings.
Correlation Analysis: • Pearson correlation coefficient is adopted.
Step 4: Analysis
L
L
Experimental Results
Sample Selection Criteria:
Count (price changes) > 50, in 6 months
Sample size:
26 out of 221 items
Experimental Results
Scaling of prices
5
1
Experimental Results
Moving Average & Tuning Parameter (window length)
Experimental Results
Shifting Analysis of prices and ratings(score)
Experimental Results
Correlation Analysis of prices and ratings
Conclusion
● Relationship exists between prices and reviews.
● Reviews influence prices in most (⅔) of the items.
● Reviews often influence prices after 7-30 days.
● Categories with loose market forces fit this rule better. o like Home, Sport, Baby
Future Work
● Improvement on sample selection.
● Analyze relationship between prices and reviews.
o For each separate category
o With an expansion from single correlation calculation
o Focus more on negative reviews
● Use our rules to predict prices.
References
[2] P. H. Calais Guerra, A. Veloso, W. Meira Jr, and V. Almeida. From bias to opinion: a transfer-learning approach to
real-time sentiment analysis. In Proceedings of the 17th ACM SIGKDD international conference on Knowledge
discovery and data mining, pages 150–158. ACM, 2011.
[3] J. L. Elsas and N. Glance. Shopping for top forums: discovering online discussion for product research. In
Proceedings of the First Workshop on Social Media Analytics, pages 23–30. ACM, 2010.
[4] M. Hu and B. Liu. Mining opinion features in customer reviews. In AAAI, volume 4, pages 755–760, 2004.
[5] J. McAuley and J. Leskovec. Hidden factors and hidden topics: understanding rating dimensions with review text. In
Proceedings of the 7th ACM conference on Recommender systems, pages 165–172. ACM, 2013.
[6] S. M. Mudambi and D. Schuff. What makes a helpful online review? a study of customer reviews on amazon.com.
Management Information Systems Quarterly, 34(1):11, 2010.
[7] B. O’Connor, R. Balasubramanyan, B. R. Routledge, and N. A. Smith. From tweets to polls: Linking text sentiment
to public opinion time series. ICWSM, 11:122–129, 2010.
[8] B. Pang, L. Lee, and S. Vaithyanathan. Thumbs up?: sentiment classification using machine learning techniques. In
Proceedings of the ACL-02 conference on Empirical methods in natural language processing-Volume 10, pages 79–86.
Association for Computational Linguistics, 2002.
[9] K. Reschke, A. Vogel, and D. Jurafsky. Generating recommendation dialogs by extracting information from user
reviews. In ACL (2), pages 499–504, 2013.
Thank you!
UVa IR Course Project Dec 5, 2014
Backup
Review Format
product/productId: B000GKXY4S
product/title: Crazy Shape Scissor Set
product/price: unknown
review/userId: A1QA985ULVCQOB
review/profileName: Carleen M. Amadio "Lady Dragonfly"
review/helpfulness: 2/2
review/score: 5.0
review/time: 1314057600
review/summary: Fun for adults too!
review/text: I really enjoy these scissors for my inspiration books that I am making (like collage, but in
books) and using these different textures these give is just wonderful, makes a great statement with
the pictures and sayings. Want more, perfect for any need you have even for gifts as well. Pretty
cool!
SNAP Amazon reviews: Products with over 100 reviews, [Aug, 2012 - Mar, 2013]
Logistics Regression Prediction Results:
Logistics Regression Prediction Results
Support Vector Machine Prediction Results:
Support Vector Machine Prediction Results