Renal echography remains the least expensive means for the exploration of the kidney. The system that we propose is a contribution for the diagnostic automatic of the kidney on ultrasound image. The analysis of texture is a technique which proved reliable in the field of the characterization of the human organs on ultrasound images. Indeed, our contribution aims at the characterization of the images of echographic textures of the kidney. This characterization is, in a first level, structural to evaluate the presence (form and position) of the various components of the kidney (clusters, medullary cortical zone). The statistical analysis of texture constitutes our second approach by carrying out virtual punctures on the kidney in order to be able to evaluate its state by quantifying the texture of the various areas characteristic of the kidney .
In the hypermedia systems the reinforcement of the learner interest requires the production, edition and diffusion of various type of teaching documents (courses, exercises, etc). The aim of our work, is the elaboration of a model of documents and teaching activities. This model describes parameters and functionalities to integrate in pedagogical contexts witch supports different activities. Based on this model, we conceived and carried out a dynamic adaptive hypermedia environment called MEDYNA, witch helps us to draft documents for e-learning. The system takes into account parameters and elements of the proposed model. It allows the dynamic generation of adaptive context to the learner. We exploited the XML technology for the implementation of our system.
This paper is concerned with the negotiation problem between agents with limited resources and under time constraints in dynamic environment. The society of agents has the same goal which is to respond with best delays at client requests. Each agent has a local technique for improving “progressively” the quality of the request. Agents must begin a negotiation cycle for coalition formation which maximizes the utility of the response to the new request. The goal is to minimize the number of exchanged messages between agents for a coalition formation in order to minimize the negotiation time.
We present in this paper a formal approach of description, posting and handling of the mathematical structured objects; based on the formalism of attribute grammars. We are interested particularly in the problem of two-dimensional and bidirectional posting of certain expressions and mathematical formulas. Indeed, in more of the two-dimensional character that presents certain mathematical symbols like the square root or the matrix, we also note the problem of posting rightto-left of an Arab text in a context planned for a posting left-to-right of an Indo-European text, or a bidirectional posting mixing the two modes. After a study of some solutions suggested in the literature, we show how the method of attribute grammars adapts easily to these types of problem.
In an earlier study it was proven and experimentally confirmed on a 2D Euler code that fixed point iterations can be differentiated to yield first and second order derivatives of implicit functions that are defined by state equations. It was also asserted that the resulting approximations for reduced gradients and Hessians converge with the same R-factor as the underlying fixed point iteration. A closer look reveals now that nevertheless these derivative values lag behind the functions in that the ratios of the corresponding errors grow proportional to the iteration counter or its square towards infinity. This rather subtle effect is caused mathematically by the occurrence of nontrivial Jordan blocks associated with degenerated eigenvalues. We elaborate the theory and report its confirmation through numerical experiments
Missing values in databases have motivated many researches in the field of KDD, specially concerning prediction. However, to the best of our knowledge, few appraoches based on association rules have been proposed so far. In this paper, we show how to adapt the levelwise algorithm for the mining of association rules in order to mine frequent rules with a confidence equal to 1 from a relational table. In our approach, the consequents of extracted rules are either an interval or a set of values, according to whether the domain of the predicted attribute is continuous or discrete.
It is well known that traditional Hidden Markov Models (HMM) systems lead to a considerable improvement when more training data or more parameters are used. However, using more data with hybrid Hidden Markov Models and Artificial Neural Networks (HMM/ANN) models results in increased training times without improvements in performance. We developed in this work a new method based on automatically separating data into several sets and training several neural networks of Multi-Layer Perceptrons (MLP) type on each set. During the recognition phase, models are combined using several criteria (based on data fusion techniques) to provide the recognized word. We showed in this paper that this method significantly improved the recognition accuracy. This method was applied in an Arabic speech recognition system. This last is based on the one hand, on a fuzzy clustering (application of the fuzzy c-means algorithm) and of another share, on a segmentation at base of the genetic algorithms.
The extremely large number of association rules that can be drawn from ―even reasonably sized datasets―, bootstrapped the development of more acute techniques or methods to reduce the size of the reported rule sets. In this context, the battery of results provided by the Formal Concept Analysis (FCA) allowed to define "irreducible" nuclei of association rule subset better known as generic basis. However, a thorough overview of the literature shows that all the algorithms dedicated neglected an essential component: the relation of order, or the extraction of the minimal generators. In this paper, we introduce the GenAll algorithm to build a formal concept lattice, in which each formal concept is "decorated" by its minimal generators. The GenAll algorithm aims to extract generic bases of association rules. The main novelty in this algorithm is the use of refinement process to compute immediate successor lists to simultaneously determine the set of formal concepts, their underlying partial order and the set of minimal generators associated with each formal concept. Carried out experiments showed that the GenAll algorithm is especially efficient for dense extraction contexts compared to that of Nourine et al. Response times obtained from the GenAll algorithm largely outperform those of Nourine et al.
GreenLab is a structural-functional model for plant growth based on multidisciplinary knowledge. Its mathematical formalism allows dynamic simulation of plant growth and model analysis. A simplified soil water balance equation is introduced to illustrate the interactions and feedbacks between the plant functioning and water resources. A water supply optimization problem is then described and solved: the sunflower fruit weight is optimized with respect to different water supply strategies in a theoretical case. Intuitive searching method and genetic algorithms are used to solve this mixed integer nonlinear problem. The optimization results are analyzed and reveal possible agronomic applications.