SQL Multimedia and Application Packages

This chapter focuses on the Structured Query Language/Multimedia (SQL/MM) and its application packages. SQL/MM is a multipart standard. Unlike SQL:1999, the various parts of SQL/MM are quite independent from one another. In late 1991 or early 1992, a specification for a language called SFQL (Structured Full-text Query Language) was issued by a group of vendors of text search engines. SFQL's stated goal was the definition of extensions to standard SQL to make that popular language more suitable for performing full-text searches of document repositories. The immediate target marketplace was the aviation industry that uses full-text systems for the management and retrieval of aircraft manufacturing and maintenance data. The proposed standard was soon to become known as SQL/MM, and the letters ‘‘MM” was understood to imply ‘‘MultiMedia.” SQL/MM Data Mining supports four different data mining techniques. One technique, the rule model, allows searching for patterns (rules) in the relationships between different parts of data. A second technique, the clustering model (sometimes called the segmentation model), helps in grouping data records that share the common characteristics and identify the most important of those characteristics. The third technique, the regression model, helps to predict the ranking of new data based on an analysis of existing data. The final technique, the classification model, shares many concepts with the regression model but is oriented toward predicting the group or class into which new data will fit based on its relationship to existing data.