High-dimensional unsupervised selection and estimation of a finite generalized Dirichlet mixture model based on minimum message length

We consider the problem of determining the structure of high-dimensional data, without prior knowledge of the number of clusters. Data are represented by a finite mixture model based on the generalized Dirichlet distribution. The generalized Dirichlet distribution has a more general covariance structure than the Dirichlet distribution and offers high flexibility and ease of use for the approximation of both symmetric and asymmetric distributions. This makes the generalized Dirichlet distribution more practical and useful. An important problem in mixture modeling is the determination of the number of clusters. Indeed, a mixture with too many or too few components may not be appropriate to approximate the true model. Here, we consider the application of the minimum message length (MML) principle to determine the number of clusters. The MML is derived so as to choose the number of clusters in the mixture model which best describes the data. A comparison with other selection criteria is performed. The validation involves synthetic data, real data clustering, and two interesting real applications: classification of web pages, and texture database summarization for efficient retrieval.     

On static analysis of tensile structures with sliding cables: the frictional sliding case

In some structural systems, such as cable structures, membranes and tensegrity structures, the use of sliding cables allows to reduce the number of elements required to be controlled during tensioning or activation. However, using sliding cables modifies the structural behavior of tensile structures since it alters the distribution of axial forces in structural members. This has been experienced in structures with continuous cables under the assumption of frictionless sliding. However, sliding-induced friction can further alter the behavior of the system. An enhancement of the static analysis of tensile structures with sliding-induced friction is investigated in this paper. In the proposed formulations, the finite-element analysis method and the dynamic relaxation method are combined with a linear complementary approach. Sliding-induced friction is integrated in the formulations through the consideration of the Euler–Eytelwein equation. The importance of considering sliding-induced friction in the static analysis of tensile structures is demonstrated through a series of examples, where it is shown that friction significantly affects the mechanical behavior of the structures. The examples also reveal that the proposed formulations do not affect the computational time of the static analyses.

     

Numerical investigation of a flexible slider–crank mechanism with multijoints with clearance

This paper aims to study the dynamic behavior of a slider–crank mechanism with flexible components and a multijoint clearance. A numerical investigation was developed for this objective. The mechanism model used for the simulation tests has been performed under MSC ADAMS software using the contact force under the “Impact-function” library. The obtained results illustrate that the mechanism performance is more significantly influenced with multiple joints with clearance. Three contact modes are involved: (i) a continuous contact motion, (ii) a free motion, and (iii) an impact motion. Numerical outcomes prove that the clearance dimension and location have a determinant effect on the slider responses precision. The mechanism reliability decay and a random overall behavior govern the dynamic response along the free flight mode.     

Parameter identification of hardening laws for bulk metal forming using experimental and numerical approach

The simulative prediction of material behaviour in forming processes necessitates a precise determination of the material parameters. The present work focusses on the modelling of the isostatic part of the flow stress using a flow curve with an analytical suppression of the influence of friction and an adequate analytical law. The experimental data are obtained from isothermal upsetting tests with various upsetting ratios. The different ratios are based on a variation of the height of the sample, remaining the diameter constant. For the proposed flow stress law five parameters are identified. In order to decrease the number of function evaluations, a new reduction model method based on both analytical and sequential quadratic programming (SQP) algorithms is developed and applied to identify flow stress law parameters. A comparison with traditional SQP algorithm is also done. A 3D finite element model is built in order to simulate a side pressing test and an experimental validation is done. As numerical results fit very well experimental data, the proposed model achieves a precise prediction of the flow behaviour. The identification of the other parts of the model (i.e. dependencies on strain-rate and temperature) are conducted in further works.     

Fatty Acids,Sterols, Polyphenols,and Chlorophylls of Olive Oils Obtained from Tunisian Wild Olive Trees (Olea europaea L. Var. Sylvestris)

Olive (Olea europaea L.) includes cultivated olive trees (var. europaea) and wild olive trees or oleaster (var. sylvestris) as two botanical varieties. These olive varieties were widely spread in the Mediterranean Region. The aim of this study was to determine fatty acid compositions, sterols, polyphenols, and chlorophylls of oils obtained from 12 wild olive trees from Northern Tunisia. Two dominated oil cultivars in Tunisia (Chétoui and Chemlali) were also used to compare results. The fatty acid methyl ester and the sterol compositions were analyzed using gas-liquid chromatography and thin layer chromatography methods, respectively. The polyphenols and chlorophylls were determined using the calorimetrical method. Results indicated that oils extracted from wild olives displayed good balanced fatty acid compositions, sterols, polyphenols, and chlorophylls. Qualitatively, for wild and cultivated olive oils, the oil has an identical composition, whereas the quantitative variation showed that some wild trees seem to be interesting oil sources as two Tunisian dominated cultivars. The highest oleic acid and polyphenol contents were 71.55% and 537.6 mg/kg of oil found in wild olives (OIch2, OIch1). The β-sitosterol was the major sterolic fraction and ranged from 84.72 to 75.70% according to the wild olives. Consequently, wild olives would be a new future edible olive oil source, as well as commonly cultivated ones.     

Single-target visual tracking using color compression and spatially weighted generalized Gaussian mixture models

Pattern Analysis and Applications - Visual tracking is a challenging task in computer vision, which intends to estimate the motion state of the target of interest in subsequent video frames. In...