Skip to Main content Skip to Navigation

Acquisition automatique et analyse de processus de décision : Application au domaine bancaire

Abstract : An essential step for the development of decision support systems is the acquisition and understanding of the decision strategies used by decision makers. The work presented in this thesis corresponds as much to a scientific problem (for the development of new techniques of expertise acquisition and for the development of cognitive based decision support systems) as to economic issues (for the development of advisory support systems in the banking domain for example). In this perspective, the APACHE software (Acquisition by Learning of Human Expert Knowledge) has been developed. It is an interactive system of automatic and real time acquisition of the individual's decision strategies. The latter, based on a cognitive model of the decision-maker (the Moving Base Heuristic), consists in making the individual classify a series of objects described by several attributes. It then produces rules, explaining the totality of the classifications made, expressed in natural language and thus easily interpretable. When decision-makers have few "time", we are subject to important constraints. The constraints concern of course the ergonomics of the software but above all the calculation time and the size of the questionnaire. Minimising the size of the training set requires in particular that the latter be generated dynamically by the algorithm (which consists in searching for anti-chains in an ordered set). We are then faced with a combinatorial optimisation problem for which we seek to minimise the calculation time. We were led to develop an original method (in terms of the algorithms implemented and the data structures) in order to minimise the acquisition time. APACHE was applied at the Crédit Mutuel de Bretagne in order to understand the attitude of individuals towards a savings product offer. The tool developed is generic and we propose other banking applications. It has also been used for face recognition judgment and for an agricultural classification problem.
Complete list of metadata
Contributor : Nathalie Fontaine Connect in order to contact the contributor
Submitted on : Tuesday, July 20, 2021 - 3:36:51 PM
Last modification on : Thursday, July 22, 2021 - 9:22:56 AM
Long-term archiving on: : Thursday, October 21, 2021 - 6:52:12 PM


Files produced by the author(s)


  • HAL Id : tel-03292746, version 1




Philippe Lenca. Acquisition automatique et analyse de processus de décision : Application au domaine bancaire. Recherche opérationnelle [cs.RO]. Université de Rennes 1, 1997. Français. ⟨NNT : 1997REN10012⟩. ⟨tel-03292746⟩



Record views


Files downloads