Results 1  10
of
1,460
An extensive empirical study of feature selection metrics for text classification
 J. of Machine Learning Research
, 2003
"... Machine learning for text classification is the cornerstone of document categorization, news filtering, document routing, and personalization. In text domains, effective feature selection is essential to make the learning task efficient and more accurate. This paper presents an empirical comparison ..."
Abstract

Cited by 483 (15 self)
 Add to MetaCart
Machine learning for text classification is the cornerstone of document categorization, news filtering, document routing, and personalization. In text domains, effective feature selection is essential to make the learning task efficient and more accurate. This paper presents an empirical comparison of twelve feature selection methods (e.g. Information Gain) evaluated on a benchmark of 229 text classification problem instances that were gathered from Reuters, TREC, OHSUMED, etc. The results are analyzed from multiple goal perspectives—accuracy, Fmeasure, precision, and recall—since each is appropriate in different situations. The results reveal that a new feature selection metric we call ‘BiNormal Separation ’ (BNS), outperformed the others by a substantial margin in most situations. This margin widened in tasks with high class skew, which is rampant in text classification problems and is particularly challenging for induction algorithms. A new evaluation methodology is offered that focuses on the needs of the data mining practitioner faced with a single dataset who seeks to choose one (or a pair of) metrics that are most likely to yield the best performance. From this perspective, BNS was the top single choice for all goals except precision, for which Information Gain yielded the best result most often. This analysis also revealed, for example, that Information Gain and ChiSquared have correlated failures, and so they work poorly together. When choosing optimal pairs of metrics for each of the four performance goals, BNS is consistently a member of the pair—e.g., for greatest recall, the pair BNS + F1measure yielded the best performance on the greatest number of tasks by a considerable margin.
<grapp@mail.utexas.edu>
"... this paper, we attempt two things: to survey the structures found in Slavic NPs from a taxonomic point of view, and b) to describe the theoretical issues involved in describing the GILBERT C. RAPPAPORT 2 ..."
Abstract
 Add to MetaCart
this paper, we attempt two things: to survey the structures found in Slavic NPs from a taxonomic point of view, and b) to describe the theoretical issues involved in describing the GILBERT C. RAPPAPORT 2
GRAPP (2011) Algarve (Portugal) Animation of air bubbles with SPH
"... We present a physicallybased multiphase model for simulating water and air bubbles with Smoothed Particle Hydrodynamics (SPH). Since the high density ratio of air and water is problematic for existing SPH solvers, we compute the density and pressure forces of both phases separately. The twoway cou ..."
Abstract
 Add to MetaCart
We present a physicallybased multiphase model for simulating water and air bubbles with Smoothed Particle Hydrodynamics (SPH). Since the high density ratio of air and water is problematic for existing SPH solvers, we compute the density and pressure forces of both phases separately. The twoway coupling is computed according to the velocity field. The proposed model is capable of simulating the complex bubble flow, e. g. path instability, deformation and merging of bubbles and volumedependent buoyancy. Furthermore, we present a velocitybased heuristic for generating bubbles in regions where air is likely trapped. Thereby, bubbles are generated on the fly, without explicitly simulating the air phase surrounding the liquid. Instead of deleting the bubbles when they reach the surface, we employ a simple foam model. By incorporating our model into the predictivecorrective SPH method, large time steps can be used. Thus, we can simulate scenarios of high resolution where the size of the bubbles is small in comparison to the liquid volume.
CLUSTERED SOLUTIONS AROUND HARMONIC CENTERS TO A COUPLED ELLIPTIC SYSTEM SOLUTIONS EN GRAPPE AUTOUR DES CENTRES HARMONIQUES
"... Abstract. We study the following system of SchrödingerMaxwell equations ε 2 ∆v − v − ωvφ + f(v) = 0, ∆φ + γv 2 = 0 in Ω, v, φ> 0 in Ω, v, φ = 0 on ∂Ω, where Ω is a smooth and bounded domain of R 3. We prove that for any integer k the system has a family of solutions (vε, φε) such that the form ..."
Abstract

Cited by 7 (0 self)
 Add to MetaCart
Abstract. We study the following system of SchrödingerMaxwell equations ε 2 ∆v − v − ωvφ + f(v) = 0, ∆φ + γv 2 = 0 in Ω, v, φ> 0 in Ω, v, φ = 0 on ∂Ω, where Ω is a smooth and bounded domain of R 3. We prove that for any integer k the system has a family of solutions (vε, φε) such that the form of vε consists of k spikes concentrating at a harmonic center of Ω as ε → 0 +. Furthermore we show that the spikes approach the vertexes of a configuration which maximizes a suitable geometrical problem. Résumé. On étudie le système d’équations de SchrödingerMaxwell suivant: ε 2 ∆v − v − ωvφ + f(v) = 0, ∆φ + γv 2 = 0 dans Ω, v, φ> 0 dans Ω, v, φ = 0 sur ∂Ω, où Ω est un ouvert borné régulier. On montre que pour tout entier k le système a une famille de solutions (vε, φε) telle que la forme de vε consiste en k pointes qui se concentrent sur un centre harmonique de Ω lorsque ε → 0 +. On montre, en plus, que les pointes approchent les sommets d’une configuration qui maximise un problème géométrique.
2011) Acceptability in interaction  From robots to Embodied Conversational Agent, Int conf on computer graphics theory and applications GRAPP 2011
"... Abstract: This paper will propose extending the traditional definition of system acceptability to take into account significant interaction between a man and a real or virtual system. We introduce the notion of “socially credible ” which describes the need of emotions and personality in intelligent ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
Abstract: This paper will propose extending the traditional definition of system acceptability to take into account significant interaction between a man and a real or virtual system. We introduce the notion of “socially credible ” which describes the need of emotions and personality in intelligent environment. This notion is measured by experiments using a plush robot Emi and a virtual embodied conversational agent 1
FlowVR: calculs interactifs et visualisation sur grappe Directeur de thèse: Mme Brigitte Plateau
"... N ◦ attribué par la bibliothèque ..."
Author manuscript, published in "International Conference on Computer Graphics Theory and Applications (GRAPP) (2009)" SIMPLE FLEXIBLE SKINNING BASED ON MANIFOLD MODELING
, 2010
"... In this paper we propose a simple framework to compute flexible skinning weights, which allows the creation from quasirigid to soft deformations. We decompose the input mesh into a set of overlapping regions, in a way similar to the constructive manifold approach. Regions are associated to skeleton ..."
Abstract
 Add to MetaCart
In this paper we propose a simple framework to compute flexible skinning weights, which allows the creation from quasirigid to soft deformations. We decompose the input mesh into a set of overlapping regions, in a way similar to the constructive manifold approach. Regions are associated to skeleton bones, and overlaps contain vertices influenced by several bones. A smooth transition function is then defined on overlaps, and is used to compute skinning weights. The size of overlaps can be tuned by the user, enabling an easy control of the desired type of deformations. 1
The competencies of regions: Canada's clusters in biotechnology
 Small Business Economics
, 2001
"... Les compétences des grappes régionales des biotechnologies au Canada ..."
Abstract

Cited by 13 (4 self)
 Add to MetaCart
Les compétences des grappes régionales des biotechnologies au Canada
Results 1  10
of
1,460