Constraint-based Decision Research Articles

Optimal constraint-based decision tree induction from itemset lattices

In this article we show that there is a strong connection between decision tree learning and local pattern mining. This connection allows us to solve the computationally hard problem of finding optimal decision trees in a wide range of applications by post-processing a set of patterns: we use local patterns to construct a global model. We exploit the connection between constraints in pattern mining and constraints in decision tree induction to develop a framework for categorizing decision tree mining constraints. This framework allows us to determine which model constraints can be pushed deeply into the pattern mining process, and allows us to improve the state-of-the-art of optimal decision tree induction.

A Constraint Based Decision Support System for Deadlocks Resolution in Collaborative New Product Design

In this paper we are considering that the product modelling step of a new product design process can be represented as a CSP (Constraint Satisfaction Problem). CSP modelling and resolution has proved its efficiency within the framework of single-designer design. We propose to extend the functions of CSP to the context of multi-designers. We have recently defined CoCSP as Cooperative Constraint Satisfaction Problem including the actors of the design problem. We present the operating principles of a new algorithm to solve the deadlocks appearing between constraints in a new product modelling process. Our reasoning is based on a model described in the form of a CoCSP. At each step, the algorithm generates the sets of consistent decisions, maximizing their cardinality. The algorithm forms the core of a prototype for an unsupervised, generic constraint-based collaborative decision support system. The work is illustrated by the example of a car engine collaborative modelling process.

BUILDING CONSISTENCIES FOR PARTIALLY DEFINED CONSTRAINTS WITH DECISION TREES AND NEURAL NETWORKS

Partially Defined Constraints can be used to model the incomplete knowledge of a concept or a relation. Instead of only computing with the known part of the constraint, we propose to complete its definition by using Machine Learning techniques. Since constraints are actively used during solving for pruning domains, building a classifier for instances is not enough: we need a solver able to reduce variable domains. Our technique is composed of two steps: first we learn a classifier for each constraint projections and then we transform the classifiers into a propagator. The first contribution is a generic meta-technique for classifier improvement showing performances comparable to boosting. The second lies in the ability of using the learned concept in constraint-based decision or optimization problems. We presents results using Decision Trees and Artificial Neural Networks for constraint learning and propagation. It opens a new way of integrating Machine Learning in Decision Support Systems.