This Turkish language dataset contains social media posts categorized into two sentiment classes. The data is distributed between 54% negative and 46% positive labels for binary classification tasks.
Use Cases
- Train a sentiment analysis model to distinguish between 'negatif' and 'pozitif' Turkish tweets
- Fine-tune Turkish BERT models for social media text classification
- Analyze the frequency of specific Turkish keywords within the 'negatif' and 'pozitif' subsets
Strengths
- Includes Turkish language text samples from Twitter
- Provides binary sentiment labels categorized as 'negatif' and 'pozitif'
- Features a balanced class distribution with 54% negative and 46% positive instances