Fouille d’épisodes à partir de séquences d’événements : Application à l’analyse de l’historique des visites de pages web

Abstract

In a constantly evolving digital environment, the analysis of user navigation on websites represents a major challenge for companies wishing to optimize their digital strategy. This dissertation explores the application of the EMDO (Episode Mining under Distinct Occurrence) algorithm to predict user behavior on the Web by analyzing their web page visit sequences. The algorithm offers an innovative approach to extracting episode-based rules based on distinct occurrences, thus improving the accuracy of predictions