200,000 labeled Korean-language movie reviews categorized into binary sentiment classes. The dataset follows the methodology of the Maas et al. (2011) Large Movie Review Dataset to provide a standardized benchmark for Korean sentiment analysis.
Use Cases
- Train a binary sentiment classifier using the 'document' text and 'label' columns
- Evaluate Korean language models on the informal text found in the 'document' field
- Develop text preprocessing pipelines for Korean morphemes using the 'document' column
Strengths
- 200,000 total samples divided into training and test sets
- Binary sentiment labels derived from Naver Movie star ratings
- Text content consists of raw Korean-language reviews scraped from the Naver platform