Abstract: Although backpropagation ANNs generally predict
better than decision trees do for pattern classification problems, they
are often regarded as black boxes, i.e., their predictions cannot be
explained as those of decision trees. In many applications, it is
desirable to extract knowledge from trained ANNs for the users to
gain a better understanding of how the networks solve the problems.
A new rule extraction algorithm, called rule extraction from artificial
neural networks (REANN) is proposed and implemented to extract
symbolic rules from ANNs. A standard three-layer feedforward ANN
is the basis of the algorithm. A four-phase training algorithm is
proposed for backpropagation learning. Explicitness of the extracted
rules is supported by comparing them to the symbolic rules generated
by other methods. Extracted rules are comparable with other methods
in terms of number of rules, average number of conditions for a rule,
and predictive accuracy. Extensive experimental studies on several
benchmarks classification problems, such as breast cancer, iris,
diabetes, and season classification problems, demonstrate the
effectiveness of the proposed approach with good generalization
ability.
Abstract: As the majority of faults are found in a few of its
modules so there is a need to investigate the modules that are
affected severely as compared to other modules and proper
maintenance need to be done in time especially for the critical
applications. As, Neural networks, which have been already applied
in software engineering applications to build reliability growth
models predict the gross change or reusability metrics. Neural
networks are non-linear sophisticated modeling techniques that are
able to model complex functions. Neural network techniques are
used when exact nature of input and outputs is not known. A key
feature is that they learn the relationship between input and output
through training. In this present work, various Neural Network Based
techniques are explored and comparative analysis is performed for
the prediction of level of need of maintenance by predicting level
severity of faults present in NASA-s public domain defect dataset.
The comparison of different algorithms is made on the basis of Mean
Absolute Error, Root Mean Square Error and Accuracy Values. It is
concluded that Generalized Regression Networks is the best
algorithm for classification of the software components into different
level of severity of impact of the faults. The algorithm can be used to
develop model that can be used for identifying modules that are
heavily affected by the faults.
Abstract: One of the most important requirements for the
operation and planning activities of an electrical utility is the
prediction of load for the next hour to several days out, known as
short term load forecasting. This paper presents the development of
an artificial neural network based short-term load forecasting model.
The model can forecast daily load profiles with a load time of one
day for next 24 hours. In this method can divide days of year with
using average temperature. Groups make according linearity rate of
curve. Ultimate forecast for each group obtain with considering
weekday and weekend. This paper investigates effects of temperature
and humidity on consuming curve. For forecasting load curve of
holidays at first forecast pick and valley and then the neural network
forecast is re-shaped with the new data. The ANN-based load models
are trained using hourly historical. Load data and daily historical
max/min temperature and humidity data. The results of testing the
system on data from Yazd utility are reported.
Abstract: The objective of this research is to examine the shear thinning behaviour of mixing flow of non-Newtonian fluid like toothpaste in the dissolution container with rotating stirrer. The problem under investigation is related to the chemical industry. Mixing of fluid is performed in a cylindrical container with rotating stirrer, where stirrer is eccentrically placed on the lid of the container. For the simulation purpose the associated motion of the fluid is considered as revolving of the container, with stick stirrer. For numerical prediction, a time-stepping finite element algorithm in a cylindrical polar coordinate system is adopted based on semi-implicit Taylor-Galerkin/pressure-correction scheme. Numerical solutions are obtained for non-Newtonian fluids employing power law model. Variations with power law index have been analysed, with respect to the flow structure and pressure drop.
Abstract: In this paper, we proposed a method to reduce
quantization error. In order to reduce quantization error, low pass
filtering is applied on neighboring samples of current block in
H.264/AVC. However, it has a weak point that low pass filtering is
performed regardless of prediction direction. Since it doesn-t consider
prediction direction, it may not reduce quantization error effectively.
Proposed method considers prediction direction for low pass filtering
and uses a threshold condition for reducing flag bit. We compare our
experimental result with conventional method in H.264/AVC and we
can achieve the average bit-rate reduction of 1.534% by applying the
proposed method. Bit-rate reduction between 0.580% and 3.567% are
shown for experimental results.
Abstract: Service quality has become a centerpiece for airline companies in vying with one another and keeps their image in the minds of passengers. Many airlines have pushed service quality through service personalization which includes both ground and on board especially from the viewpoint of retaining satisfied passengers and attracting new ones. Besides those, in-flight meals/food service is another important aspect of the airline operation. The in flight meals/food services now are seen as part of marketing strategies in attracting business or leisure travelers. This study reports the outcomes of the investigation on in-flight meals/food attributes toward passengers- level of satisfaction and re-flying intention. Taste, freshness, appearance of in-flight meals/food served and menu choices are important to the airlines passengers especially for the long haul flight. Food not only contributes to the prediction of the airline passengers- levels of satisfaction but besides other factors slightly influence passengers- re- flying intention. Airline companies therefore should not ignore this element but take the opportunity to create more attractive and acceptable in-flight meals/food along with other matter as marketing tools in attracting passengers to re-flying with them.
Abstract: In order to make conventional implicit algorithm to be applicable in large scale parallel computers , an interface prediction and correction of discontinuous finite element method is presented to solve time-dependent neutron transport equations under 2-D cylindrical geometry. Domain decomposition is adopted in the computational domain.The numerical experiments show that our parallel algorithm with explicit prediction and implicit correction has good precision, parallelism and simplicity. Especially, it can reach perfect speedup even on hundreds of processors for large-scale problems.
Abstract: Determination of wellbore problems during a
production/injection process might be evaluated thorough
temperature log analysis. Other applications of this kind of log
analysis may also include evaluation of fluid distribution analysis
along the wellbore and identification of anomalies encountered
during production/injection process. While the accuracy of such
prediction is paramount, the common method of determination of a
wellbore temperature log includes use of steady-state energy balance
equations, which hardly describe the real conditions as observed in
typical oil and gas flowing wells during production operation; and
thus increase level of uncertainties. In this study, a practical method
has been proposed through development of a simplified semianalytical
model to apply for predicting temperature profile along the
wellbore. The developed model includes an overall heat transfer
coefficient accounting all modes of heat transferring mechanism,
which has been focused on the prediction of a temperature profile as
a function of depth for the injection/production wells. The model has
been validated with the results obtained from numerical simulation.
Abstract: Post cracking behavior and load –bearing capacity of
the steel fiber reinforced high-strength concrete (SFRHSC) are
dependent on the number of fibers are crossing the weakest crack
(bridged the crack) and their orientation to the crack surface. Filling
the mould by SFRHSC, fibers are moving and rotating with the
concrete matrix flow till the motion stops in each internal point of the
concrete body. Filling the same mould from the different ends
SFRHSC samples with the different internal structures (and different
strength) can be obtained. Numerical flow simulations (using Newton
and Bingham flow models) were realized, as well as single fiber
planar motion and rotation numerical and experimental investigation
(in viscous flow) was performed. X-ray pictures for prismatic
samples were obtained and internal fiber positions and orientations
were analyzed. Similarly fiber positions and orientations in cracked
cross-section were recognized and were compared with numerically
simulated. Structural SFRHSC fracture model was created based on
single fiber pull-out laws, which were determined experimentally.
Model predictions were validated by 15x15x60cm prisms 4 point
bending tests.
Abstract: Tandem mass spectrometry (MS/MS) is the engine
driving high-throughput protein identification. Protein mixtures possibly
representing thousands of proteins from multiple species are
treated with proteolytic enzymes, cutting the proteins into smaller
peptides that are then analyzed generating MS/MS spectra. The
task of determining the identity of the peptide from its spectrum
is currently the weak point in the process. Current approaches to de
novo sequencing are able to compute candidate peptides efficiently.
The problem lies in the limitations of current scoring functions. In this
paper we introduce the concept of proteome signature. By examining
proteins and compiling proteome signatures (amino acid usage) it is
possible to characterize likely combinations of amino acids and better
distinguish between candidate peptides. Our results strongly support
the hypothesis that a scoring function that considers amino acid usage
patterns is better able to distinguish between candidate peptides. This
in turn leads to higher accuracy in peptide prediction.
Abstract: This paper presents the development of a hybrid
thermal model for the EVO Electric AFM 140 Axial Flux Permanent
Magnet (AFPM) machine as used in hybrid and electric vehicles. The
adopted approach is based on a hybrid lumped parameter and finite
difference method. The proposed method divides each motor
component into regular elements which are connected together in a
thermal resistance network representing all the physical connections
in all three dimensions. The element shape and size are chosen
according to the component geometry to ensure consistency. The
fluid domain is lumped into one region with averaged heat transfer
parameters connecting it to the solid domain. Some model parameters
are obtained from Computation Fluid Dynamic (CFD) simulation and
empirical data. The hybrid thermal model is described by a set of
coupled linear first order differential equations which is discretised
and solved iteratively to obtain the temperature profile. The
computation involved is low and thus the model is suitable for
transient temperature predictions. The maximum error in temperature
prediction is 3.4% and the mean error is consistently lower than the
mean error due to uncertainty in measurements. The details of the
model development, temperature predictions and suggestions for
design improvements are presented in this paper.
Abstract: The main goal of data mining is to extract accurate, comprehensible and interesting knowledge from databases that may be considered as large search spaces. In this paper, a new, efficient type of Genetic Algorithm (GA) called uniform two-level GA is proposed as a search strategy to discover truly interesting, high-level prediction rules, a difficult problem and relatively little researched, rather than discovering classification knowledge as usual in the literatures. The proposed method uses the advantage of uniform population method and addresses the task of generalized rule induction that can be regarded as a generalization of the task of classification. Although the task of generalized rule induction requires a lot of computations, which is usually not satisfied with the normal algorithms, it was demonstrated that this method increased the performance of GAs and rapidly found interesting rules.
Abstract: The present study focuses on methods allowing a convenient and quick calculation of the SIFs in order to predict the static adhesive strength of bonded joints. A new SIF calculation method is proposed, based on the stresses obtained from a FE model at a reference point located in the adhesive layer at equal distance of the free-edge and of the two interfaces. It is shown that, even limiting ourselves to the two main modes, i.e. the opening and the shearing modes, and using the values of the stresses resulting from a low detailed FE model, an efficient calculation of the peeling stress at adhesive-substrate corners can be obtained by this way. The proposed method is interesting in that it can be the basis of a prediction tool that will allow the designer to quickly evaluate the SIFs characterizing a particular application without developing a detailed analysis.
Abstract: The quantified residence time distribution (RTD)
provides a numerical characterization of mixing in a reactor, thus
allowing the process engineer to better understand mixing
performance of the reactor.This paper discusses computational
studies to investigate flow patterns in a two impinging streams
cyclone reactor(TISCR) . Flow in the reactor was modeled with
computational fluid dynamics (CFD). Utilizing the Eulerian-
Lagrangian approach, implemented in FLUENT (V6.3.22), particle
trajectories were obtained by solving the particle force balance
equations. From simulation results obtained at different Δts, the mean
residence time (tm) and the mean square deviation (σ2) were
calculated. a good agreement can be observed between predicted and
experimental data. Simulation results indicate that the behavior of
complex reactor systems can be predicted using the CFD technique
with minimum data requirement for validation.
Abstract: As it is known, buoyancy and drag forces rule bubble's rise velocity in a liquid column. These forces are strongly dependent on fluid properties, gravity as well as equivalent's diameter. This study reports a set of bubble rising velocity experiments in a liquid column using water or glycerol. Several records of terminal velocity were obtained. The results show that bubble's rise terminal velocity is strongly dependent on dynamic viscosity effect. The data set allowed to have some terminal velocities data interval of 8.0 ? 32.9 cm/s with Reynolds number interval 1.3 -7490. The bubble's movement was recorded with a video camera. The main goal is to present an original set data and results that will be discussed based on two-phase flow's theory. It will also discussed, the prediction of terminal velocity of a single bubble in liquid, as well as the range of its applicability. In conclusion, this study presents general expressions for the determination of the terminal velocity of isolated gas bubbles of a Reynolds number range, when the fluid proprieties are known.
Abstract: A multiple-option analytical model for the evaluation of the energy performance and distribution of aerodynamic forces acting on a vertical-axis Darrieus wind turbine depending on both rotor architecture and operating conditions is presented. For this purpose, a numerical algorithm, capable of generating the desired rotor conformation depending on design geometric parameters, is coupled to a Single/Double-Disk Multiple-Streamtube Blade Element – Momentum code. Both single and double-disk configurations are analyzed and model predictions are compared to literature experimental data in order to test the capability of the code for predicting rotor performance. Effective airfoil characteristics based on local blade Reynolds number are obtained through interpolation of literature low-Reynolds airfoil databases. Some corrections are introduced inside the original model with the aim of simulating also the effects of blade dynamic stall, rotor streamtube expansion and blade finite aspect ratio, for which a new empirical relationship to better fit the experimental data is proposed. In order to predict also open field rotor operation, a freestream wind shear profile is implemented, reproducing the effect of atmospheric boundary layer.
Abstract: In this research a comparison between k-epsilon and
LES model for a deoiling hydrocyclone is conducted. Flow field of
hydrocyclone is obtained by three-dimensional simulations with
OpenFOAM code. Potential of prediction for both methods of this
complex swirl flow is discussed. Large eddy simulation method
results have more similarity to experiment and its results are
presented in figures from different hydrocyclone cross sections.
Abstract: A multi-agent system is developed here to predict
monthly details of the upcoming peak of the 24th solar magnetic
cycle. While studies typically predict the timing and magnitude of
cycle peaks using annual data, this one utilizes the unsmoothed
monthly sunspot number instead. Monthly numbers display more
pronounced fluctuations during periods of strong solar magnetic
activity than the annual sunspot numbers. Because strong magnetic
activities may cause significant economic damages, predicting
monthly variations should provide different and perhaps helpful
information for decision-making purposes. The multi-agent system
developed here operates in two stages. In the first, it produces twelve
predictions of the monthly numbers. In the second, it uses those
predictions to deliver a final forecast. Acting as expert agents, genetic
programming and neural networks produce the twelve fits and
forecasts as well as the final forecast. According to the results
obtained, the next peak is predicted to be 156 and is expected to
occur in October 2011- with an average of 136 for that year.
Abstract: There is lot of work done in prediction of the fault proneness of the software systems. But, it is the severity of the faults that is more important than number of faults existing in the developed system as the major faults matters most for a developer and those major faults needs immediate attention. In this paper, we tried to predict the level of impact of the existing faults in software systems. Neuro-Fuzzy based predictor models is applied NASA-s public domain defect dataset coded in C programming language. As Correlation-based Feature Selection (CFS) evaluates the worth of a subset of attributes by considering the individual predictive ability of each feature along with the degree of redundancy between them. So, CFS is used for the selecting the best metrics that have highly correlated with level of severity of faults. The results are compared with the prediction results of Logistic Models (LMT) that was earlier quoted as the best technique in [17]. The results are recorded in terms of Accuracy, Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). The results show that Neuro-fuzzy based model provide a relatively better prediction accuracy as compared to other models and hence, can be used for the modeling of the level of impact of faults in function based systems.
Abstract: Uncertainties of a serial production line affect on the
production throughput. The uncertainties cannot be prevented in a
real production line. However the uncertain conditions can be
controlled by a robust prediction model. Thus, a hybrid model
including autoregressive integrated moving average (ARIMA) and
multiple polynomial regression, is proposed to model the nonlinear
relationship of production uncertainties with throughput. The
uncertainties under consideration of this study are demand, breaktime,
scrap, and lead-time. The nonlinear relationship of production
uncertainties with throughput are examined in the form of quadratic
and cubic regression models, where the adjusted R-squared for
quadratic and cubic regressions was 98.3% and 98.2%. We optimized
the multiple quadratic regression (MQR) by considering the time
series trend of the uncertainties using ARIMA model. Finally the
hybrid model of ARIMA and MQR is formulated by better adjusted
R-squared, which is 98.9%.