The rapid growth of the internet has increased people's reliance on it for expressing opinions on products and stores. Text sentiment analysis is now a key research area. Deep learning methods are commonly used for text classification, but they suffer from issues like information loss and weak context. This paper enhances the existing models to simplify the process, educe training time, and improve overall recall and accuracy in text sentiment classification. With the rapid development of artificial intelligence, traditional manual techniques are fading, and AI algorithms are driving the swift progress of text sentiment classification. Product categorization is critical in e-commerce since it influences search, recommendations, inventory, and consumer experience. NLP approaches and models such as TF-IDF and Word2Vec are effective for categorizing products based on their textual descriptions. This literature study investigates the research evolution, problems, and upcoming trends in using TF-IDF and Word2Vec for product classification in e-commerce.