Identifying and Handling Mislabelled Instances

Data mining and knowledge discovery aim at producing useful and reliable models from the data. Unfortunately some databases contain noisy data which perturb the generalization of the models. An important source of noise consists of mislabelled training instances. We offer a new approach which deals with improving classification accuracies by using a preliminary filtering procedure. An example is suspect when in its neighbourhood defined by a geometrical graph the proportion of examples of the same class is not significantly greater than in the database itself. Such suspect examples in the training data can be removed or relabelled. The filtered training set is then provided as input to learning algorithms. Our experiments on ten benchmarks of UCI Machine Learning Repository using 1-NN as the final algorithm show that removal gives better results than relabelling. Removing allows maintaining the generalization error rate when we introduce from 0 to 20% of noise on the class, especially when classes are well separable. The filtering method proposed is finally compared to the relaxation relabelling schema.     

Optimizing NFV placement for distributing micro-data centers in cellular networks

With the popularity of mobile devices, the next generation of mobile networks has faced several challenges. Different applications have been emerged, with different requirements. Offering an infrastructure that meets different types of applications with specific requirements is one of these issues. In addition, due to user mobility, the traffic generated by the mobile devices in a specific location is not constant, making it difficult to reach the optimal resource allocation. In this context, network function virtualization (NFV) can be used to deploy the telecommunication stacks as virtual functions running on commodity hardware to meet users’ requirements such as performance and availability. However, the deployment of virtual functions can be a complex task. To select the best placement strategy that reduces the resource usage, at the same time keeps the performance and availability of network functions is a complex task, already proven to be an NP-hard problem. Therefore, in this paper, we formulate the NFV placement as a multi-objective problem, where the risk associated with the placement and energy consumption are taken into consideration. We propose the usage of two optimization algorithms, NSGA-II and GDE3, to solve this problem. These algorithms were taken into consideration because both work with multi-objective problems and present good performance. We consider a triathlon circuit scenario based on real data from the Ironman route as an use case to evaluate and compare the algorithms. The results show that GDE3 is able to attend both objectives (minimize failure and minimize energy consumption), while the NSGA-II prioritizes energy consumption.

     

Beneficial effect of gamma irradiation on the N-deacetylation of chitin to form chitosan

The effect of gamma irradiation on the N-deacetylation of chitin to form chitosan was studied. Chitin from crab shells was irradiated up to 20 kGy and N-deacetylated in aqueous NaOH solution (40% and 60% w/w) at 60 and 100 °C for 60 min. The degree of N-deacetylation (DD) of non-irradiated and irradiated samples was determined by IR-band ratio method. It was found that higher extent of N-deacetylation was achieved for the chitin samples irradiated up to 20 kGy doses as compared to non-irradiated chitin. The DD values of chitin, prepared from non-irradiated and 20 kGy irradiated chitins by N-deacetylation at 60 °C with 40% NaOH for 60 min, were found to be 38% and 60%, respectively. The increase in DD by irradiation was interpreted as a result of reduction in molecular weight of chitin. Low dose irradiation of chitin has provided the possibility of its N-deacetylation into chitosan at much milder reaction conditions.     

Numerical simulation of exchange membrane fuel cells in different operating conditions

A two-dimensional, steady state model for proton exchange membrane fuel cell (PEMFC) is presented. The model is used to describe the effect operation conditions (current density, pressure and water content) on the water transport, ohmic resistance and water distribution in the membrane and performance of PEMFC. This model considers the transport of species and water along the porous media: gas diffusion layers (GDL) anode and cathode, and the membrane of PEMFC fuel cell.     

Transport phenomena effect on the performance of proton exchange membrane fuel cell (PEMFC)

The aim of this work is to present a two-dimensional transient model, of heat and mass transfer in a proton exchange membrane fuel cell (PEMFC). The model includes various conservation equations such as mass (hydrogen, oxygen, water concentration), Momentum and energy equations this model is combined with the electrochemical model.     

One dimensional transient numerical study of the mass heat and charge transfer in a proton exchange membrane for PEMFC

The objective of our study is to quantify the mass water transferred by various modes: diffusion, convection and migration. For the water transfer, the principal forces considered in the model are, the convection force, the osmotic force (i.e. diffusion) and the electric force (migration). The first of these forces results from a pressure gradient, the second of a concentration gradient and the third of a protons' migration from the anode to the cathode, which has an effect on the dipole of the water molecules (resistance force to the advancement). The numerical tool used to solve the equations' system is the finite element method. The results obtained numerically considering this method are concentration profiles and concentration variation with time and membrane thickness. These results illustrate the contribution of each mass transfer mode.     

Discrete Distortion in Triangulated 3‐Manifolds

We introduce a novel notion, that we call discrete distortion, for a triangulated 3‐manifold. Discrete distortion naturally generalizes the notion of concentrated curvature defined for triangulated surfaces and provides a powerful tool to understand the local geometry and topology of 3‐manifolds. Discrete distortion can be viewed as a discrete approach to Ricci curvature for singular flat manifolds. We distinguish between two kinds of distortion, namely, vertex distortion, which is associated with the vertices of the tetrahedral mesh decomposing the 3‐manifold, and bond distortion, which is associated with the edges of the tetrahedral mesh. We investigate properties of vertex and bond distortions. As an example, we visualize vertex distortion on manifold hypersurfaces in R⁴ defined by a scalar field on a 3D mesh. distance fields.