地球资源数据云——数据资源详情
电信客户流失 + 真实的客户反馈
该数据集《Telco Customer Churn + Realistic Customer Feedback》主要用于多分类任务，数据形态以图像为主，应用场景偏向文本内容分析。 题目说明：Enriched version with 7000+ AI - generated customer reviews, ready for NLP 任务类型：图像多分类。 建议流程：先检查类别分布与脏样本，再用迁移学习（如 ResNet/EfficientNet）建立基线。 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：telco_churn_with_all_feedback.csv, telco_noisy_feedback_prep.csv, telco_prep.csv。 This dataset extends the classic Telco Customer Churn dataset by adding a column CustomerFeedback, generated using GPT - 3.5 based on customer features like tenure, contract type, monthly charges, and churn status. Each row now includes a realistic customer review, enabling: Sentiment analysis NLP model training (classification, clustering) End - to - end churn prediction pipelines (ML + NLP)
摘要概览
该数据集《Telco Customer Churn + Realistic Customer Feedback》主要用于多分类任务，数据形态以图像为主，应用场景偏向文本内容分析。 题目说明：Enriched version with 7000+ AI - generated customer reviews, ready for NLP
任务类型：图像多分类。
建议流程：先检查类别分布与脏样本，再用迁移学习（如 ResNet/EfficientNet）建立基线。
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：telco_churn_with_all_feedback.csv, telco_noisy_feedback_prep.csv, telco_prep.csv。
This dataset extends the classic Telco Customer Churn dataset by adding a column CustomerFeedback, generated using GPT - 3.5 based on customer features like tenure, contract type, monthly charges, and churn status.
Each row now includes a realistic customer review, enabling:
Sentiment analysis
NLP model training (classification, clustering)
End - to - end churn prediction pipelines (ML + NLP)