RecSys_Dataset: Beauty Product reviews dataset for sentiment analysis and recommendation system

Product reviews help the sellers to understand their customers' expectations and sentiment towards the product and based on those reviews they take measures accordingly to heighten the satisfaction level of their customers. Beauty products are unique because various factors can influence a customer's purchase decision. With the help of machine learning techniques, the product reviews can be utilized to achieve insights and patterns to understand customers sentiment and recommend products according to their purchase records. To maintain the confidentiality of user, real dataset was not used. A synthetic dataset can heighten the efficiency of machine learning techniques. This dataset was generated by AI, packs a vast number of reviews of various products, sentiment towards those and elaborate exploratory analysis. A total of 50,000 reviews were generated from 200 different products and 1,000 unique users. A series of processing steps were performed on the raw dataset, including content addition. Another aspect of this work is that there are still not many datasets available that contains user CTR (Click-Through Rate) alongside their Spent time on a product interface which can help the researchers exploit this dataset to develop recommender systems, and natural language processing algorithms for analytical purposes.商品评论可帮助卖家洞悉消费者对产品的期待与情感倾向，卖家据此采取对应举措以提升消费者的满意度。美妆个护产品具有其特殊性，诸多因素均可影响消费者的购买决策。借助机器学习技术，可对商品评论进行挖掘以获取洞察与模式，进而洞悉消费者情感倾向，并基于消费者的购买记录为其推荐产品。为保护用户隐私，本研究未使用真实数据集。合成数据集可提升机器学习技术的应用效率。本数据集由人工智能生成，涵盖海量多品类商品评论及其情感倾向，并附带详尽的探索性分析内容。该数据集共生成50000条评论，涉及200款不同产品与1000位独立用户。研究团队对原始数据集执行了一系列处理流程，其中包括内容增补环节。本研究的另一特色在于，当前兼具用户点击通过率（Click-Through Rate, CTR）与用户在产品界面停留时长数据的数据集仍较为稀缺，这使得本数据集可助力研究人员开发推荐系统与自然语言处理算法以开展相关分析工作。