1 / 12

Great Food, Lousy Service

Great Food, Lousy Service. Topic Modeling for Sentiment Analysis in Sparse Reviews. Robin Melnick rmelnick@stanford.edu. Dan Preston dpreston@stanford.edu. OpenTable.com. Short. Words. Characters. Sparse.

sonel
Download Presentation

Great Food, Lousy Service

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Great Food, Lousy Service Topic Modeling for Sentiment Analysis in Sparse Reviews Robin Melnick rmelnick@stanford.edu Dan Preston dpreston@stanford.edu

  2. OpenTable.com

  3. Short Words Characters

  4. Sparse “An unexpected combination of Left-Bank Paris and Lower Manhattan in Omaha. Divine. Inspirational and a great value.” • Food? • Ambiance? • Service? • Noise?

  5. Skewed

  6. Correlations

  7. SVM + Features, Features, Features! • 30+ preprocessing and SVM classification features, • ~50 configurations

  8. Key Features • Stemming • Porter 1980 via NLTK • <fast>, <faster>, <fastest> <fast> • Negation processing • (enhanced approach from Pang et al. 2002) • “Not a great experience.”  NOT_great • “They neverdisappoint!”  NOT_disappoint • Net sentiment count • pos/neg lexicon (Harvard General Inquirer) • running +/- count • “Incredible(+) food, but our server was rude(-).”  (0)

  9. Results (so far) • Trained on 10,000 reviews • Tested on ~80,000 reviews • Accuracy • Baseline: 50.0% • Intermediate model: 56.6% (1.13x) • abs( average scoring delta ): 0.56

  10. Topic Modeling Hand-seeded topic-word list expanded via WordNetSynSets • sub-topic classifiers • topic-filtered n-grams • <soupFOOD was fantasticADJ> • <fantasticADJsoupFOOD was> • topic-word proximity filtering • both above  <fantasticADJ/FOOD>. Results:

  11. Word-Rating Distributions “decent” “worst” “mediocre” “solid” “exceeded”

  12. Frequency-Weighted Entropy Model • Accuracy • Baseline: 50.0% • Intermediate model: 56.6% • Best (entropy) model: 58.6% (1.17x) • abs( average scoring delta ): 0.56 0.52

More Related