Statistical methods for data mining and knowledge discovery

Jean Vaillancourt

pp. 51-60

in: Lonard Kwuida, Bar Sertkaya (eds), Formal concept analysis, Berlin, Springer, 2010

Abstract

This survey paper aims mainly at giving computer scientists a rapid bird's eye view, from a mathematician's perspective, of the main statistical methods used in order to extract knowledge from databases comprising various types of observations. After touching briefly upon the matters of supervision, data regularization and a brief review of the main models, the key issues of model assessment, selection and inference are perused. Finally, specific statistical problems arising from applications around data mining and warehousing are explored. Examples and applications are chosen mainly from the vast collection of image and video retrieval, indexation and classification challenges facing us today.