ABSTRACT
This chapter overviews several specialized data mining techniques. It introduces time-series analysis and shows how neural networks are used to solve time-series problems. The chapter shows how data mining is used for website evaluation, personalization, and adaptation. It also provides an overview of how the PageRank algorithm is used to determine the prestige of a page or website. The chapter offers an overview of textual data mining. It discusses methods for dealing with large-sized data, imbalanced data, and streaming data. The chapter presents two methods that, in some cases, can improve the classification correctness of supervised learner models. Multiple-model methods such as bagging and boosting can sometimes improve model performance. Both approaches work best with unstable data mining algorithms. Data mining can be applied to help determine perceptions users have about the websites they visit. The data available as a result of one or more Web-based user sessions are stored in Web server log files.
