Résumé. Ces dernières années, la classification croisée ou classification par blocs, c’est-à-dire la recherche simultanée d’une partition des lignes et d’une partition des colonnes d’un tableau de données, est devenue un outil très utilisé en fouille de données. Dans ce domaine, l’information se présente souvent sous forme de tableaux de contingence ou tableaux de co-occurrence croisant les mod...

Clustering techniques have always been oriented to solve classification and pattern recognition problems. This clustering techniques have been used also to initialize the centers of the Radial Basis Function (RBF) when designing an RBF Neural Network (RBFNN) that approximates a function. Since classification and function approximation problems are quite different, it is necessary to design a ne...

A granular computing model is used for learning classification rules by considering the two basic issues: concept formation and concept relationships identification. A classification rule induction method is proposed. Instead of focusing on the selection of a suitable partition, i.e., a family of granules defined by values of an attribute, in each step, we concentrate on the selection of a sing...

When clustering algorithms are applied to image seg-mentation, the goal is to solve a classification problem. However, these algorithms do not directly optimize classification quality. As a result, they are susceptible to two problems: P1) the criterion they optimize may not be a good estimator of " true " classification quality, and P2) they often admit many (suboptha€) solutions. This paper i...

Due to the high dimensionality of a hyperspectral image, classification accuracy of a single classifier may be limited when the size of the training set is small. A divide-and-conquer approach has been proposed, where a classifier is applied to each group of bands and the final output will be the fused result of multiple classifiers. Since the dimensionality in each band group is much lower, cl...

We introduce a new network planning approach for management of multicast communications in large data dissemination networks. Our approach, addressing Channelization Problem, considers both the common user interests and geographical dependencies while evaluating multicast configurations. We first perform Global Similarity Classification to partition users into virtual domains based on their cha...

Multi-modal classification problems arise in many fields and form an important class of problems. The presence of disjoint areas for each class creates special problems for techniques that cannot partition each class into more than one region. Among the various techniques that have been applied with some success to multi-modal problems are decision tree classifiers (DTCs) and back propagation n...

We propose a multi-partition, multi-chunk ensemble classifier based data mining technique to classify concept-drifting data streams. Existing ensemble techniques in classifying concept-drifting data streams follow a single-partition, single-chunk approach, in which a single data chunk is used to train one classifier. In our approach, we train a collection of v classifiers from r consecutive dat...

several techniques exist to select and materialize a proper set of data in a suitable structure that manage the queries submitted to the online analytical processing systems. these techniques are called view management techniques, which consist of three research areas: 1) view selection to materialize, 2) query processing and rewriting using the materialized views, and 3) maintaining materializ...

Thematic maps of spatial data are constructed by using standard thematic classification methods that do not allow management the uncertainty and, consequently, evaluation reliability resulting map. We propose a novel fuzzy-based method applied to construct in Geographical Information Systems. An initial fuzzy partition domain features dataset is triangular numbers; our finds an optimal evaluati...