Paper: My Curiosity was Satisfied, but not in a Good Way: Predicting User Ratings for Online Recipes

ACL ID W14-5903
Title My Curiosity was Satisfied, but not in a Good Way: Predicting User Ratings for Online Recipes
Venue Workshop on Natural Language Processing for Social Media
Session  
Year 2014
Authors

In this paper, we develop an approach to automatically predict user ratings for recipes at Epicuri- ous.com, based on the recipes? reviews. We investigate two distributional methods for feature se- lection, Information Gain and Bi-Normal Separation; we also compare distributionally selected features to linguistically motivated features and two types of frameworks: a one-layer system where we aggregate all reviews and predict the rating vs. a two-layer system where ratings of individual reviews are predicted and then aggregated. We obtain our best results by using the two-layer architecture, in combination with 5 000 features selected by Information Gain. This setup reaches an overall accuracy of 65.60%, given an upper bound of 82.57%.

@InProceedings{liu-EtAl:2014:SocialNLP,
  author    = {Liu, Can  and  Guo, Chun  and  Dakota, Daniel  and  Rajagopalan, Sridhar  and  Li, Wen  and  K\"{u}bler, Sandra  and  Yu, Ning},
  title     = {"My Curiosity was Satisfied, but not in a Good Way": Predicting User Ratings for Online Recipes},
  booktitle = {Proceedings of the Second Workshop on Natural Language Processing for Social Media (SocialNLP)},
  month     = {August},
  year      = {2014},
  address   = {Dublin, Ireland},
  publisher = {Association for Computational Linguistics and Dublin City University},
  pages     = {12--21},
  url       = {http://www.aclweb.org/anthology/W14-5903}
}