Users' behavior in a web browser can be categorized into two states, namely, the search state and the browse state.
Four types of data summarization that are widely used in log mining, namely, query histograms, click-through bipartites, click patterns, and session patterns.
Silvestri divided log mining technologies into two
major categories. The rst category focuses on enhancing the eciency of a search
system, while the second category focuses on enhancing the eectiveness of a system.
From the viewpoint of eectiveness, a web
search system usually consists of three basic components, namely, query understand-
ing, document understanding, and document ranking, and one optional component,
namely user understanding.