This document discusses bag-of-words methods for text mining applications such as document retrieval, opinion mining, and association mining. It describes how a basic bag-of-words approach can be used for document retrieval by creating topic vectors from word frequencies and calculating cosine similarity between documents. Opinion mining can be done by training a naive Bayes classifier on labeled documents to learn positive and negative words. Association mining seeks relationships between terms, but a simple bag-of-words approach has limitations and better methods use term classification and learning from contexts.