A bag-of-words model represents text by counting or indicating which words appear, usually without preserving grammar or word order.
The approach is simple and useful for classification, retrieval, and baseline NLP work, but it loses context that modern embeddings and transformer models can represent more effectively.