Learn More
Variable subset selection is an important issue when learning regression models from data. In this paper, a filter approach to subset variable selection based on the notion of mutual information is compared with a wrapper approach based on fast non-parametric regression algorithms. A detailed comparison of the two approaches is made on three nonlinear and(More)
In this paper, we present a new way to model agent and multi-agents systems (Mas). The modeling we introduce relies on the proposal of a MAS generic model, and a hybrid model of aptitudes supporting the specific agent's properties (autonomy, independence, adaptation). In addition to the formal description of the agent's particular features, the hybrid model(More)
Two main modelisation techniques are the analytic and systemic methods. After a presentation of each method applied to building simulation, this paper emphases the benefit of systemic modelisation for multizone cases. Using a conceptual description of a building issued of previous method, one particular subsystem appears naturally, i.e. the thermal zones.(More)
OBJECTIVE Data about endocrine and bone disease in nontransfusion-dependent thalassaemia (NTDT) is scanty. The aim of our study was to evaluate these complications in β-TI adult patients. METHODS We studied retrospectively 70 β-TI patients with mean followup of 20 years. Data recorded included age, gender, haemoglobin and ferritin levels, biochemical and(More)
  • Emmanuelle Sisavath, Aude Mazuel, Stephan J. Jorry, Nathalie Babonneau, Patrick Bachèlery, Béatrice de Voogd +5 others
  • 2013
The present study focused on turbidite sedimentation in the Cilaos turbidite system, a volcaniclastic deepsea fan recently recognized offshore La Réunion Island. A set of piston cores was collected in order to establish the stratigraphy of this fan and to examine the processes controlling the turbidite sedimentation off the Cilaos cirque (Piton des Neiges(More)
The first and maybe the most important step in designing a model-based predictive controller is to develop a model that is as accurate as possible and that is valid under a wide range of operating conditions. The sugar boiling process is a strongly nonlinear and nonstationary process. The main process nonlinearities are represented by the crystal growth(More)
We present a way to specify a Galois k weak hierarchy of a meet-description context , using, on the one hand, weak clusters associated with a multiway dissimilarity function and, on the other hand, a strict valuation on the entity description space. Moreover, we place ourselves in the framework of a probabilistic meet-closed description context, i.e., a(More)
When multi-agent research is used in electronic commerce, most of time, the study relates to the negotiation. The general idea is to send agents negotiating in place of human negotiator, after preliminary allotted the appropriate behavior. The behavior depends on the type of transaction. The negotiation is however only one angle of vision. We should not(More)
  • 1