ebrukilic/ytu-ara-proje-absa

ABSA数据集是一个用于土耳其语基于方面的情感分析（ABSA）的数据集，包含5045个实例。数据集涵盖了多个产品类别（如裙子、外套、衬衫、毛衣、裤子）和情感极性（负面、中性、正面）。数据来源于Trendyol和HepsiBurada网站上的服装产品评论，经过预处理和标准化处理。数据集以CSV格式提供，每个评论都有唯一的ID、星级评分、产品子类别、标准化评论文本、涉及的方面（如质量、面料、颜色、尺寸、物流、价格）和情感极性。数据集的创建者还提到了可能的偏见和局限性，如标签可能仅代表特定用户群体，且数据集仅涵盖特定类型的产品。

The ABSA (Aspect-Based Sentiment Analysis) dataset contains 5045 instances in Turkish, with the task of aspect-based sentiment analysis. The dataset categories include skirts, coats, shirts, sweaters, and pants, with sentiment polarity divided into negative, neutral, and positive, and aspects including quality, fabric, color, size, shipping, and price. The dataset is sourced from clothing product reviews on the Trendyol and HepsiBurada websites, in CSV format, with comments preprocessed and normalized. The dataset was created by students from Yıldız Teknik Üniversitesi Bilgisayar Mühendisliği, with domain experts @rmysysv and @ebrukilic. The dataset annotations were completed by the dataset creators. The dataset may have biases and limitations, such as the dataset may not represent all user groups and only includes specific types of products.