Abstract: As in today's semiconductor industries test costs can make up to 50 percent of the total production costs, an efficient test error detection becomes more and more important. In this paper, we present a new machine learning approach to test error detection that should provide a faster recognition of test system faults as well as an improved test error recall. The key idea is to learn a classifier ensemble, detecting typical test error patterns in wafer test results immediately after finishing these tests. Since test error detection has not yet been discussed in the machine learning community, we define central problem-relevant terms and provide an analysis of important domain properties. Finally, we present comparative studies reflecting the failure detection performance of three individual classifiers and three ensemble methods based upon them. As base classifiers we chose a decision tree learner, a support vector machine and a Bayesian network, while the compared ensemble methods were simple and weighted majority vote as well as stacking. For the evaluation, we used cross validation and a specially designed practical simulation. By implementing our approach in a semiconductor test department for the observation of two products, we proofed its practical applicability.
Abstract: This paper explores the effectiveness of machine
learning techniques in detecting firms that issue fraudulent financial
statements (FFS) and deals with the identification of factors
associated to FFS. To this end, a number of experiments have been
conducted using representative learning algorithms, which were
trained using a data set of 164 fraud and non-fraud Greek firms in the
recent period 2001-2002. The decision of which particular method to
choose is a complicated problem. A good alternative to choosing
only one method is to create a hybrid forecasting system
incorporating a number of possible solution methods as components
(an ensemble of classifiers). For this purpose, we have implemented
a hybrid decision support system that combines the representative
algorithms using a stacking variant methodology and achieves better
performance than any examined simple and ensemble method. To
sum up, this study indicates that the investigation of financial
information can be used in the identification of FFS and underline the
importance of financial ratios.
Abstract: Magnetic and semiconductor nanomaterials exhibit
novel magnetic and optical properties owing to their unique size and
shape-dependent effects. With shrinking the size down to nanoscale
region, various anomalous properties that normally not present in bulk
start to dominate. Ability in harnessing of these anomalous properties
for the design of various advance electronic devices is strictly
dependent on synthetic strategies. Hence, current research has focused
on developing a rational synthetic control to produce high quality
nanocrystals by using organometallic approach to tune both size and
shape of the nanomaterials. In order to elucidate the growth
mechanism, transmission electron microscopy was employed as a
powerful tool in performing real time-resolved morphologies and
structural characterization of magnetic (Fe3O4) and semiconductor
(ZnO) nanocrystals. The current synthetic approach is found able to
produce nanostructures with well-defined shapes. We have found that
oleic acid is an effective capping ligand in preparing oxide-based
nanostructures without any agglomerations, even at high temperature.
The oleate-based precursors and capping ligands are fatty acid
compounds, which are respectively originated from natural palm oil
with low toxicity. In comparison with other synthetic approaches in
producing nanostructures, current synthetic method offers an effective
route to produce oxide-based nanomaterials with well-defined shapes
and good monodispersity. The nanocystals are well-separated with
each other without any stacking effect. In addition, the as-synthesized
nanopellets are stable in terms of chemically and physically if
compared to those nanomaterials that are previous reported. Further
development and extension of current synthetic strategy are being
pursued to combine both of these materials into nanocomposite form
that will be used as “smart magnetic nanophotocatalyst" for industry
waste water treatment.
Abstract: The governing two-dimensional equations of a heterogeneous material composed of a fluid (allowed to flow in the absence of acoustic excitations) and a crystalline piezoelectric cubic solid stacked one-dimensionally (along the z direction) are derived and special emphasis is given to the discussion of acoustic group velocity for the structure as a function of the wavenumber component perpendicular to the stacking direction (being the x axis). Variations in physical parameters with y are neglected assuming infinite material homogeneity along the y direction and the flow velocity is assumed to be directed along the x direction. In the first part of the paper, the governing set of differential equations are derived as well as the imposed boundary conditions. Solutions are provided using Hamilton-s equations for the wavenumber vs. frequency as a function of the number and thickness of solid layers and fluid layers in cases with and without flow (also the case of a position-dependent flow in the fluid layer is considered). In the first part of the paper, emphasis is given to the small-frequency case. Boundary conditions at the bottom and top parts of the full structure are left unspecified in the general solution but examples are provided for the case where these are subject to rigid-wall conditions (Neumann boundary conditions in the acoustic pressure). In the second part of the paper, emphasis is given to the general case of larger frequencies and wavenumber-frequency bandstructure formation. A wavenumber condition for an arbitrary set of consecutive solid and fluid layers, involving four propagating waves in each solid region, is obtained again using the monodromy matrix method. Case examples are finally discussed.
Abstract: This paper deals with a numerical analysis of the
transient response of composite beams with strain rate dependent
mechanical properties by use of a finite difference method. The
equations of motion based on Timoshenko beam theory are derived.
The geometric nonlinearity effects are taken into account with von
Kármán large deflection theory. The finite difference method in
conjunction with Newmark average acceleration method is applied to
solve the differential equations. A modified progressive damage
model which accounts for strain rate effects is developed based on
the material property degradation rules and modified Hashin-type
failure criteria and added to the finite difference model. The
components of the model are implemented into a computer code in
Mathematica 6. Glass/epoxy laminated composite beams with
constant and strain rate dependent mechanical properties under
dynamic load are analyzed. Effects of strain rate on dynamic
response of the beam for various stacking sequences, load and
boundary conditions are investigated.
Abstract: The present study deals with the analysis of the cylindrical part of a CNG storage vessel, combining a plastic liner and an over wrapped filament wound composite. Three kind of polymer are used in the present analysis: High density Polyethylene HDPE, Light low density Polyethylene LLDPE and finally blend of LLDPE/HDPE. The effect of the mechanical properties on the behavior of type IV vessel may be then investigated. In the present paper, the effect of the order of the circumferential winding on the stacking sequence may be then investigated. Based on mechanical considerations, the present model provides an exact solution for stresses and deformations on the cylindrical section of the vessel under thermo-mechanical static loading. The result show a good behavior of HDPE liner compared to the other plastic materials. The presence of circumferential winding angle in the stacking improves the rigidity of vessel by improving the burst pressure.