Term Statistics for Structured Text Retrieval

Classical ranking algorithms in information retrieval make use of term statistics, the most common (and basic) ones being within-document term frequency, tf, and document frequency, df. tf is the number of occurrences of a term in a document and is used to reflect how well a term captures the topic of a document, whereas df is the number of documents in which a term appears and is used to reflect how well a term discriminates between relevant and non-relevant documents. df is also commonly referred to as inverse document frequency, idf, since it is inversely related to the importance of a term. Both tf and idf are obtained at indexing time. Ranking algorithms for structured text retrieval, and more precisely XML retrieval, require similar terms statistics, but with respect to elements.