Thesis : Exploring sequential data with Relational Concept Analysis

PhD defense : Cristina NICA

Team : SDC

Title : Exploring sequential data with Relational Concept Analysis

Abstract : Nowadays, large amounts of sequential data are generated and stored in order to be further harnessed by discovering valuable pieces of information. Many sequential pattern mining methods have been proposed to discover potentially useful and understandable patterns that describe the analysed sequential data. These works have focused on efficiently enumerating all the patterns or concise representations, such as closed partially-ordered patterns (cpo-patterns), that makes their evaluation a laboured task for domain experts since their number can be quite large. To address this issue, we propose a new approach, that is to directly extract multilevel cpo-patterns implicitly organised into a hierarchy. To this end, we devise an original and self-contained method within the Relational Concept Analysis (RCA) framework, referred to as RCA-SEQ, that exploits the relational nature of sequential data and the structure and properties of the lattices from the RCA output. RCA-SEQ spans five steps: (1) the preprocessing of the raw data; (2) the RCA-based exploration of the preprocessed data; (3) the automatic extraction of a hierarchy of multilevel cpo-patterns by navigating the lattices from the RCA output; (4) the selection of relevant multilevel cpo-patterns based on various measures of interest; (5) the pattern evaluation done by domain experts. In addition, we show that the RCA-SEQ approach can be easily adapted to extract more informative patterns (the weighted cpo-patterns), to integrate a user-defined taxonomy or to explore heterogeneous sequential data. Two hydro-ecological datasets have been used to asses RCA-SEQ.

This thesis was supervised by Mrs LE BER Florence, Ingénieure en chef des Ponts, des Eaux et des Forêts, HDR, ENGEES, University of Strasbourg, ICube.

The jury was composed by Mr PONCELET Pascal Professor (Montpellier University, LIRMM), Mr SOLDANO Henry Associate prodessor HDR (Paris-North University, LIPN), Mr FERRÉ Sébastien Associate professor HDR (Rennes 1 University, IRISA), Mr BEISEL Jean-Nicolas Professor (ENGEES, Strasbourg University), Mme BRAUD Agnès Associate professor (Strasbourg University, ICube).

Cette thèse a été dirigée par Mme LE BER Florence, Ingénieure en chef des Ponts, des Eaux et des Forêts, HDR, ENGEES, Université de Strasbourg, ICube.

The PhD defense (in english) will be held on Friday, 13th october 2017 at 3.00pm in the A302 amphitheater of the pole API building in Illkirch.