撰写Aspect Based Sentiment Analysis论文，寻求二元情感分类Yelp数据集

Hey there! I get you're working on an Aspect-Based Sentiment Analysis (ABSA) paper and need binary (positive/negative) Yelp datasets used in those two papers—let's break down how you can track them down:

• For the 2017 conference paper

  • Reach out to the paper's corresponding author directly: Most academic papers list contact info for the lead/corresponding author. Shoot them a polite email explaining your ABSA research focus (binary sentiment comparison experiments) and request access to their dataset. Researchers usually share data with fellow academics for legitimate work.
  • Check for supplementary materials: Some conference papers include appendices or linked supplementary files with dataset access details. Double-check the paper's page for any hidden links or notes about data availability.
  • Explore the conference's dataset archive: The conference proceedings (vol143) might have a dedicated section for shared datasets from presented papers—browse through the event's official pages to see if this dataset is hosted there.

• For the Springer chapter

  • Contact the chapter authors: You can find author contact details on the chapter's page. A clear, concise email outlining your research goal and need for their binary Yelp dataset is your best bet here.
  • Look for supplementary materials: Springer often allows authors to attach supplementary data to their chapters. Check the chapter's page for a "Supplementary Materials" tab or section that might link to the dataset.
  • Build your own aligned dataset from Yelp's official open data: Yelp has a public open dataset with reviews labeled with star ratings. You can filter for 1-2 star (negative) and 4-5 star (positive) reviews, then apply the aspect extraction/annotation rules described in the two papers to replicate their dataset. This is a great fallback if you can't get direct access, as it ensures your data matches the papers' preprocessing logic.

A quick pro tip: When emailing authors, be sure to mention your affiliation and specific research use case—it helps build trust and makes them more likely to share the data.

内容的提问来源于stack exchange，提问作者M. Erfan Mowlaei