A heterogeneous fuzzy decision framework for extremity bias in online reviews

Dataset Description This dataset supports the research article titled "A heterogeneous fuzzy decision framework for extremity bias in online reviews." The raw data were collected from dianping.com using the Octopus collector. All data are publicly available web content, and no private or personally identifiable information was included. Data content The dataset contains CSV and XLSX files. Variables include: 1. review_text: original user comment text 2. review_rating: star rating Preprocessing Data cleaning procedures: remove duplicate records, filter invalid empty reviews, and eliminate abnormal outlier samples. Text preprocessing: Chinese word segmentation and stop word removal using jieba. The annotated subset was manually coded by 3 coders. Inter-coder reliability was calculated using Cohen's Kappa. File list - Excel: raw collected data and preprocessed review, etc - results: comparative analysis results - readme.txt: variable dictionary - code: Python scripts for preprocessing the review dataset, transforming it into N-IVPLTS, and comparative analysis Usage Notes This dataset can be used for academic research only. Users should cite this Zenodo record and the original paper when using the data.