Featuring crawled records of submissions to the ICLR 2023 conference, encompassing paper titles, abstracts, and author lists. It tracks the peer review process through numerical ratings, reviewer confidence scores, and final acceptance decisions for thousands of entries.
Use Cases
- Predict the 'decision' outcome of a paper using text features from the 'abstract' and 'title' columns.
- Calculate the variance in 'rating' scores per paper to measure reviewer disagreement levels.
- Extract research trends by performing topic modeling on the 'keywords' and 'abstract' fields.
- Analyze the impact of 'confidence' scores on the final 'decision' using regression analysis.
Strengths
- Captures submission metadata including 'title', 'abstract', and 'keywords' for ICLR 2023 papers.
- Includes quantitative 'rating' and 'confidence' scores provided by individual reviewers.
- Records the final 'decision' status (e.g., Accept, Reject) for each manuscript.
- Maps 'author' names and affiliations to their respective research submissions.