The chapter introduces the concept of automatic keyword extraction from individual documents, emphasizing its importance in information retrieval (IR) systems. Keywords are defined as sequences of words that compactly represent the content of a document, aiding in query definition, document indexing, and retrieval. The chapter highlights the limitations of manual keyword assignment and the need for automated methods. Early approaches to automated keyword extraction focused on corpus-oriented statistics, but these methods often fail to capture context and can be limited by the corpus size. Document-oriented methods, which operate on individual documents, are more context-independent and can scale to large collections. These methods typically combine natural language processing (NLP) techniques with machine learning algorithms to identify and score candidate keywords based on co-occurrence patterns. RAKE (Rapid Automatic Keyword Extraction) is introduced as an unsupervised, domain-independent method for extracting keywords from individual documents. It uses a set of stop words and phrase delimiters to partition the document into candidate keywords, which are then scored based on their co-occurrence patterns. RAKE is designed to be efficient, simple, and effective, outperforming existing methods in terms of precision and computational efficiency. The chapter also discusses the evaluation of RAKE using a benchmark dataset of technical abstracts, demonstrating its superior performance compared to TextRank and supervised learning methods. Additionally, it introduces a method for generating stoplists, which can be tailored to specific corpora and domains, further enhancing the effectiveness of RAKE. Finally, the chapter evaluates RAKE's performance on a corpus of news articles, MPQA Corpus, to assess its ability to extract essential and general keywords. The results show that RAKE can effectively identify keywords that are essential or general to the corpus, making it a valuable tool for various applications in IR and text analysis.The chapter introduces the concept of automatic keyword extraction from individual documents, emphasizing its importance in information retrieval (IR) systems. Keywords are defined as sequences of words that compactly represent the content of a document, aiding in query definition, document indexing, and retrieval. The chapter highlights the limitations of manual keyword assignment and the need for automated methods. Early approaches to automated keyword extraction focused on corpus-oriented statistics, but these methods often fail to capture context and can be limited by the corpus size. Document-oriented methods, which operate on individual documents, are more context-independent and can scale to large collections. These methods typically combine natural language processing (NLP) techniques with machine learning algorithms to identify and score candidate keywords based on co-occurrence patterns. RAKE (Rapid Automatic Keyword Extraction) is introduced as an unsupervised, domain-independent method for extracting keywords from individual documents. It uses a set of stop words and phrase delimiters to partition the document into candidate keywords, which are then scored based on their co-occurrence patterns. RAKE is designed to be efficient, simple, and effective, outperforming existing methods in terms of precision and computational efficiency. The chapter also discusses the evaluation of RAKE using a benchmark dataset of technical abstracts, demonstrating its superior performance compared to TextRank and supervised learning methods. Additionally, it introduces a method for generating stoplists, which can be tailored to specific corpora and domains, further enhancing the effectiveness of RAKE. Finally, the chapter evaluates RAKE's performance on a corpus of news articles, MPQA Corpus, to assess its ability to extract essential and general keywords. The results show that RAKE can effectively identify keywords that are essential or general to the corpus, making it a valuable tool for various applications in IR and text analysis.