Pascal Francq

Graph nodes clustering based on the commute-time kernel

This work presents a kernel method for clustering the nodes of a weighted, undirected, graph. The algorithm is based on a two-step procedure. First, the sigmoid commute-time kernel (KCT), providing a similarity measure between any couple of nodes by taking the indirect links into account, is computed from the adjacency matrix of the graph. Then, the nodes of the graph are clustered by performing a kernel kmeans or fuzzy k-means on this CT kernel matrix. For this purpose, a new, simple, version of the kernel k-means and the kernel fuzzy k-means is introduced. The joint use of the CT kernel matrix and kernel clustering appears to be quite successful. Indeed, it provides good results on a document clustering problem involving the newsgroups database.

Can Web 2.0 Applications Save e-Democracy? A Study of How New Internet Applications May Enhance Citizen Participation in the Political Process Online

Increased citizen participation is a core element of both Web 2.0 applications and the concept(s) of e-democracy. This paper proposes a prospective view of how the Web 2.0 can be used in the context of e-democracy. After a review of both concepts, the main perceptions of internet users concerning their political activity on the internet are highlighted by an online survey realised in January-March 2007. Various Web 2.0 applications are then discussed in the light of VEDELs axis – information, discussion and decision-making – in order to identify how they may enhance the implication of citizens in the political process through the internet.

XML components ranking : Which relevant ranking criteria ? Which relevant criteria merging ?

Retrieving specific information becomes an important issue, especially in XML information retrieval (XML-IR). Indexing XML components (elements) rather than the entire XML document allows to target more effectively the relevant elements. In this paper, we consider the problem of ranking the relevant components returned in response to a user query. As opposed to ranking documents, ranking XML elements introduces new criteria to take into account in determining the degree of relevance of a component, as for example the granularity of an XML component or the proximity of query keywords within the XML structure. Our aim is to answer two questions : Which criteria reflect best the relevance of an XML component? and how can these criteria be combined to obtain an efficient scoring function in order to calculate the degree of relevance of an XML component?

An adapted genetic algorithm to solve generalized cell formation

Abstract This paper addresses the cell formation problem with alternative process plans and machine capacity constraints. Given alternative process plans, machine capacities and quantities of parts to produce, the problem consists in defining preferential process and routing for each part (grouping of operations into machines) optimizing machines grouping into manufacturing cells. The problem can be decomposed in two distinct sub-problems: operations grouping on resources, yielding flows between the machines, and grouping of these latter into independent cells. The objective of the proposed method is to optimize both groupings (operations on machines and machines into cell) minimizing of the inter-cellular moves. To solve simultaneously both grouping interdependent problems, we propose a modified grouping genetic algorithm (SIGGA). In our adapted grouping genetic algorithm, each chromosome is composed of two parts, one part for each problem. According to different application rates, the genetic operators are applied on the first, on the second problem, or on both problems. Finally, the population chromosomes simultaneously evolve in both problems.