Date: Dec. 8, 2001
Link: http://www.cs.rpi.edu/~puninj/LOGML/
Source Author or Organization: John Punin, Mukkai Krishnamoorthy, Gerard Uffelman, Department of Computer Science, Rensselaer Polytechnic Institute (RPI)

A great deal of information about visitors to Web sites is stored in various computer logs, often more than the visitor or even site owners or maintainers are aware of. Information that can be deduced from or massaged out of Web server log files is valuable to marketers and other dater miners. Log Markup Language (LOGML) is an XML-based Web log report description language for commercial and research Web data mining applications.

LOGML uses Extensible Graph Markup and Modeling Language (XGMML) and a framework the developers call Frequent Pattern Mining to analyze Web Usage Mining data and generate a graph of user visits information, summary reports of information such as browser type and usage time statistics and a sub-graph detailing user activity while on a site. LOGML can be combined with Resource Description Framework (RDF) to provide metadata information, Extensible Style Language Transformations (XSLT) to output HTML reports and Scalable Vector Graphics (SVG) to illustrate reports.

The LOGML site provides links to the LOGML Draft Specification, LOGML Document Type Definition (DTD) and LOGML W3C XML Schema.