Due to the outbreak of the epidemic in recent years, the development of e-commerce has led to a golden age. However, since customers can't personally articulate their product dissatisfaction or satisfaction to the store, the review section has become a place where customers often discuss or share reviews that help other new customers understand the true condition of each store's products. And e-commerce product reviews are emotionally featured, from which we can extract key emotionally featured words to drive customers' attitudes toward the product. In this study, we will use the Vader function to extract keywords, and then the use polarity score to judge women customers' recommendation index for women's clothing e-commerce platform, and analyze which aspects are the most important for e-commerce store owners to pay attention to and improve in order to convert the bad reviews