地球资源数据云——数据资源详情
阿拉伯语公司评论 (查看更多)
该数据集《Arabic Company Reviews (عربي)》主要用于多分类任务，数据形态以文本为主，应用场景偏向文本内容分析。 题目说明：40K+ reviews in Arabic for sentiment analysis (NLP) 任务类型：文本多分类。 建议流程：先做文本清洗与分词，再比较 TF - IDF+线性模型 与 预训练语言模型。 评估建议：使用分层切分或交叉验证，优先关注 F1、Recall、AUC 等分类指标。 可用文件：CompanyReviews.csv。 Context The data was collected for the purpose of sentiment analysis in order to produce a score for a given company. Content The data has 40K+ reviews in Arabic for sentiment analysis each labelled with a rating and its associated company name. Dataset Glossary (Column - Wise)
摘要概览
该数据集《Arabic Company Reviews (عربي)》主要用于多分类任务，数据形态以文本为主，应用场景偏向文本内容分析。 题目说明：40K+ reviews in Arabic for sentiment analysis (NLP)
任务类型：文本多分类。
建议流程：先做文本清洗与分词，再比较 TF - IDF+线性模型 与 预训练语言模型。
评估建议：使用分层切分或交叉验证，优先关注 F1、Recall、AUC 等分类指标。
可用文件：CompanyReviews.csv。
Context
The data was collected for the purpose of sentiment analysis in order to produce a score for a given company.
Content
The data has 40K+ reviews in Arabic for sentiment analysis each labelled with a rating and its associated company name.
Dataset Glossary (Column - Wise)