Efficiently Finding Conceptual Clustering Models with Integer Linear Programming

Authors: Abdelkader Ouali, Samir Loudni, Yahia Lebbah, Patrice Boizumault, Albrecht Zimmermann, Lakhdar Loukil

IJCAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments performed on UCI datasets show that our approach efficiently finds clusterings of consistently high quality.
Researcher Affiliation Academia Abdelkader Oualia,b, Samir Loudnib, Yahia Lebbaha, Patrice Boizumaultb, Albrecht Zimmermannb and Lakhdar Loukila (a) University of Oran 1 Ahmed Ben Bella, Lab. LITIO, 31000 Oran, Algeria. (b) University of Caen Normandy, CNRS, UMR 6072 GREYC, 14032 Caen, France.
Pseudocode No The paper provides ILP model formulations in figures (Fig. 1, 2, 3) but these are mathematical problem definitions, not pseudocode or algorithm blocks.
Open Source Code No The paper provides a link for a re-implemented third-party tool (CDKMeans) but does not state that the code for their own proposed method (CCLP) is available or open-source.
Open Datasets Yes Experiments were carried out on the same datasets which were used in [Guns et al., 2013] and available from the UCI repository.
Dataset Splits No The paper refers to datasets from the UCI repository and lists their characteristics in Table 3, but does not provide specific train/validation/test dataset splits for reproducibility.
Hardware Specification Yes All experiments were conducted on AMD Opteron 6282SE with 2.60 GHz of CPU and 512 GB of RAM.
Software Dependencies Yes We used LCM to extract the set of all closed patterns and CPLEX v.12.4 to solve the different ILP models.
Experiment Setup Yes For all methods, a time limit of 24 hours has been used. Experiments have been performed without any local constraints on individual closed patterns. To evaluate the quality of a clustering, we test the coherence of a clustering, measured by the intra-cluster similarity (ICS) and the inter-clusters dissimilarity (ICD), both of which should be as large as possible.