Abstract: In this paper back-propagation artificial neural network
(BPANN )with Levenberg–Marquardt algorithm is employed to
predict the deformation of the upsetting process. To prepare a
training set for BPANN, some finite element simulations were
carried out. The input data for the artificial neural network are a set
of parameters generated randomly (aspect ratio d/h, material
properties, temperature and coefficient of friction). The output data
are the coefficient of polynomial that fitted on barreling curves.
Neural network was trained using barreling curves generated by
finite element simulations of the upsetting and the corresponding
material parameters. This technique was tested for three different
specimens and can be successfully employed to predict the
deformation of the upsetting process
Abstract: In this paper, a neural tree (NT) classifier having a
simple perceptron at each node is considered. A new concept for
making a balanced tree is applied in the learning algorithm of the
tree. At each node, if the perceptron classification is not accurate and
unbalanced, then it is replaced by a new perceptron. This separates
the training set in such a way that almost the equal number of patterns
fall into each of the classes. Moreover, each perceptron is trained only
for the classes which are present at respective node and ignore other
classes. Splitting nodes are employed into the neural tree architecture
to divide the training set when the current perceptron node repeats
the same classification of the parent node. A new error function based
on the depth of the tree is introduced to reduce the computational
time for the training of a perceptron. Experiments are performed to
check the efficiency and encouraging results are obtained in terms of
accuracy and computational costs.
Abstract: With increasing complexity in electronic systems
there is a need for system level anomaly detection and fault isolation.
Anomaly detection based on vector similarity to a training set is used
in this paper through two approaches, one the preserves the original
information, Mahalanobis Distance (MD), and the other that
compresses the data into its principal components, Projection Pursuit
Analysis. These methods have been used to detect deviations in
system performance from normal operation and for critical parameter
isolation in multivariate environments. The study evaluates the
detection capability of each approach on a set of test data with known
faults against a baseline set of data representative of such “healthy"
systems.
Abstract: In this work, statistical experimental design was
applied for the optimization of medium constituents for Gentamicin
production by Micromsonospora echinospora subs pallida (MTCC
708) in a batch reactor and the results are compared with the ANN
predicted values. By central composite design, 50 experiments are
carried out for five test variables: Starch, Soya bean meal, K2HPO4,
CaCO3 and FeSO4. The optimum condition was found to be: Starch
(8.9,g/L), Soya bean meal (3.3 g/L), K2HPO4 (0.8 g/L), CaCO3 (4
g/L) and FeSO4 (0.03 g/L). At these optimized conditions, the yield
of gentamicin was found to be 1020 mg/L. The R2 values were found
to be 1 for ANN training set, 0.9953 for ANN test set, and 0.9286 for
RSM.
Abstract: Recordings from recent earthquakes have provided evidence that ground motions in the near field of a rupturing fault differ from ordinary ground motions, as they can contain a large energy, or “directivity" pulse. This pulse can cause considerable damage during an earthquake, especially to structures with natural periods close to those of the pulse. Failures of modern engineered structures observed within the near-fault region in recent earthquakes have revealed the vulnerability of existing RC buildings against pulse-type ground motions. This may be due to the fact that these modern structures had been designed primarily using the design spectra of available standards, which have been developed using stochastic processes with relatively long duration that characterizes more distant ground motions. Many recently designed and constructed buildings may therefore require strengthening in order to perform well when subjected to near-fault ground motions. Fiber Reinforced Polymers are considered to be a viable alternative, due to their relatively easy and quick installation, low life cycle costs and zero maintenance requirements. The objective of this paper is to investigate the adequacy of Artificial Neural Networks (ANN) to determine the three dimensional dynamic response of FRP strengthened RC buildings under the near-fault ground motions. For this purpose, one ANN model is proposed to estimate the base shear force, base bending moments and roof displacement of buildings in two directions. A training set of 168 and a validation set of 21 buildings are produced from FEA analysis results of the dynamic response of RC buildings under the near-fault earthquakes. It is demonstrated that the neural network based approach is highly successful in determining the response.
Abstract: Human activity is a major concern in a wide variety of
applications, such as video surveillance, human computer interface
and face image database management. Detecting and recognizing
faces is a crucial step in these applications. Furthermore, major
advancements and initiatives in security applications in the past years
have propelled face recognition technology into the spotlight. The
performance of existing face recognition systems declines significantly
if the resolution of the face image falls below a certain level.
This is especially critical in surveillance imagery where often, due to
many reasons, only low-resolution video of faces is available. If these
low-resolution images are passed to a face recognition system, the
performance is usually unacceptable. Hence, resolution plays a key
role in face recognition systems. In this paper we introduce a new
low resolution face recognition system based on mixture of expert
neural networks. In order to produce the low resolution input images
we down-sampled the 48 × 48 ORL images to 12 × 12 ones using
the nearest neighbor interpolation method and after that applying
the bicubic interpolation method yields enhanced images which is
given to the Principal Component Analysis feature extractor system.
Comparison with some of the most related methods indicates that
the proposed novel model yields excellent recognition rate in low
resolution face recognition that is the recognition rate of 100% for
the training set and 96.5% for the test set.
Abstract: This paper introduces a hand gesture recognition system to recognize real time gesture in unstrained environments. Efforts should be made to adapt computers to our natural means of communication: Speech and body language. A simple and fast algorithm using orientation histograms will be developed. It will recognize a subset of MAL static hand gestures. A pattern recognition system will be using a transforrn that converts an image into a feature vector, which will be compared with the feature vectors of a training set of gestures. The final system will be Perceptron implementation in MATLAB. This paper includes experiments of 33 hand postures and discusses the results. Experiments shows that the system can achieve a 90% recognition average rate and is suitable for real time applications.
Abstract: The one-class support vector machine “support vector
data description” (SVDD) is an ideal approach for anomaly or outlier
detection. However, for the applicability of SVDD in real-world
applications, the ease of use is crucial. The results of SVDD are
massively determined by the choice of the regularisation parameter C
and the kernel parameter of the widely used RBF kernel. While for
two-class SVMs the parameters can be tuned using cross-validation
based on the confusion matrix, for a one-class SVM this is not
possible, because only true positives and false negatives can occur
during training. This paper proposes an approach to find the optimal
set of parameters for SVDD solely based on a training set from
one class and without any user parameterisation. Results on artificial
and real data sets are presented, underpinning the usefulness of the
approach.
Abstract: Much research into handwritten Thai character
recognition have been proposed, such as comparing heads of
characters, Fuzzy logic and structure trees, etc. This paper presents a
system of handwritten Thai character recognition, which is based on
the Ant-minor algorithm (data mining based on Ant colony
optimization). Zoning is initially used to determine each character.
Then three distinct features (also called attributes) of each character
in each zone are extracted. The attributes are Head zone, End point,
and Feature code. All attributes are used for construct the
classification rules by an Ant-miner algorithm in order to classify
112 Thai characters. For this experiment, the Ant-miner algorithm is
adapted, with a small change to increase the recognition rate. The
result of this experiment is a 97% recognition rate of the training set
(11200 characters) and 82.7% recognition rate of unseen data test
(22400 characters).
Abstract: In the automotive industry test drives are being conducted
during the development of new vehicle models or as a part of
quality assurance of series-production vehicles. The communication
on the in-vehicle network, data from external sensors, or internal
data from the electronic control units is recorded by automotive
data loggers during the test drives. The recordings are used for fault
analysis. Since the resulting data volume is tremendous, manually
analysing each recording in great detail is not feasible.
This paper proposes to use machine learning to support domainexperts
by preventing them from contemplating irrelevant data and
rather pointing them to the relevant parts in the recordings. The
underlying idea is to learn the normal behaviour from available
recordings, i.e. a training set, and then to autonomously detect
unexpected deviations and report them as anomalies.
The one-class support vector machine “support vector data description”
is utilised to calculate distances of feature vectors. SVDDSUBSEQ
is proposed as a novel approach, allowing to classify subsequences
in multivariate time series data. The approach allows to
detect unexpected faults without modelling effort as is shown with
experimental results on recordings from test drives.
Abstract: In determining the electromagnetic properties of
magnetic materials, hysteresis modeling is of high importance. Many
models are available to investigate those characteristics but they tend
to be complex and difficult to implement. In this paper a new
qualitative hysteresis model for ferromagnetic core is presented,
based on the function approximation capabilities of adaptive neuro
fuzzy inference system (ANFIS). The proposed ANFIS model
combined the neural network adaptive capabilities and the fuzzy
logic qualitative approach can restored the hysteresis curve with a
little RMS error. The model accuracy is good and can be easily
adapted to the requirements of the application by extending or
reducing the network training set and thus the required amount of
measurement data.
Abstract: Support vector machines (SVMs) are considered to be
the best machine learning algorithms for minimizing the predictive
probability of misclassification. However, their drawback is that for
large data sets the computation of the optimal decision boundary is a
time consuming function of the size of the training set. Hence several
methods have been proposed to speed up the SVM algorithm. Here
three methods used to speed up the computation of the SVM
classifiers are compared experimentally using a musical genre
classification problem. The simplest method pre-selects a random
sample of the data before the application of the SVM algorithm. Two
additional methods use proximity graphs to pre-select data that are
near the decision boundary. One uses k-Nearest Neighbor graphs and
the other Relative Neighborhood Graphs to accomplish the task.
Abstract: The quick training algorithms and accurate solution
procedure for incremental learning aim at improving the efficiency of
training of SVR, whereas there are some disadvantages for them, i.e.
the nonconvergence of the formers for changeable training set and
the inefficiency of the latter for a massive dataset. In order to handle
the problems, a new training algorithm for a changeable training
set, named Approximation Incremental Training Algorithm (AITA),
was proposed. This paper explored the reason of nonconvergence
theoretically and discussed the realization of AITA, and finally
demonstrated the benefits of AITA both on precision and efficiency.
Abstract: In this paper back-propagation artificial neural network
(BPANN) is employed to predict the deformation of the upsetting
process. To prepare a training set for BPANN, some finite element
simulations were carried out. The input data for the artificial neural
network are a set of parameters generated randomly (aspect ratio d/h,
material properties, temperature and coefficient of friction). The
output data are the coefficient of polynomial that fitted on barreling
curves. Neural network was trained using barreling curves generated
by finite element simulations of the upsetting and the corresponding
material parameters. This technique was tested for three different
specimens and can be successfully employed to predict the
deformation of the upsetting process
Abstract: There have been significant improvements in automatic
voice recognition technology. However, existing systems still face difficulties,
particularly when used by non-native speakers with accents.
In this paper we address a problem of identifying the English accented
speech of speakers from different backgrounds. Once an accent is
identified the speech recognition software can utilise training set from
appropriate accent and therefore improve the efficiency and accuracy
of the speech recognition system. We introduced the Q factor, which
is defined by the sum of relationships between frequencies of the
formants. Four different accents were considered and experimented
for this research. A scoring method was introduced in order to
effectively analyse accents. The proposed concept indicates that the
accent could be identified by analysing their formants.
Abstract: Surface roughness (Ra) is one of the most important requirements in machining process. In order to obtain better surface roughness, the proper setting of cutting parameters is crucial before the process take place. This research presents the development of mathematical model for surface roughness prediction before milling process in order to evaluate the fitness of machining parameters; spindle speed, feed rate and depth of cut. 84 samples were run in this study by using FANUC CNC Milling α-Τ14ιE. Those samples were randomly divided into two data sets- the training sets (m=60) and testing sets(m=24). ANOVA analysis showed that at least one of the population regression coefficients was not zero. Multiple Regression Method was used to determine the correlation between a criterion variable and a combination of predictor variables. It was established that the surface roughness is most influenced by the feed rate. By using Multiple Regression Method equation, the average percentage deviation of the testing set was 9.8% and 9.7% for training data set. This showed that the statistical model could predict the surface roughness with about 90.2% accuracy of the testing data set and 90.3% accuracy of the training data set.
Abstract: In this paper back-propagation artificial neural
network (BPANN) with Levenberg–Marquardt algorithm is
employed to predict the limiting drawing ratio (LDR) of the deep
drawing process. To prepare a training set for BPANN, some finite
element simulations were carried out. die and punch radius, die arc
radius, friction coefficient, thickness, yield strength of sheet and
strain hardening exponent were used as the input data and the LDR
as the specified output used in the training of neural network. As a
result of the specified parameters, the program will be able to
estimate the LDR for any new given condition. Comparing FEM and
BPANN results, an acceptable correlation was found.
Abstract: Text Mining is around applying knowledge discovery techniques to unstructured text is termed knowledge discovery in text (KDT), or Text data mining or Text Mining. In Neural Network that address classification problems, training set, testing set, learning rate are considered as key tasks. That is collection of input/output patterns that are used to train the network and used to assess the network performance, set the rate of adjustments. This paper describes a proposed back propagation neural net classifier that performs cross validation for original Neural Network. In order to reduce the optimization of classification accuracy, training time. The feasibility the benefits of the proposed approach are demonstrated by means of five data sets like contact-lenses, cpu, weather symbolic, Weather, labor-nega-data. It is shown that , compared to exiting neural network, the training time is reduced by more than 10 times faster when the dataset is larger than CPU or the network has many hidden units while accuracy ('percent correct') was the same for all datasets but contact-lences, which is the only one with missing attributes. For contact-lences the accuracy with Proposed Neural Network was in average around 0.3 % less than with the original Neural Network. This algorithm is independent of specify data sets so that many ideas and solutions can be transferred to other classifier paradigms.
Abstract: Segmentation techniques based on Active Contour
Models have been strongly benefited from the use of prior information
during their evolution. Shape prior information is captured from
a training set and is introduced in the optimization procedure to
restrict the evolution into allowable shapes. In this way, the evolution
converges onto regions even with weak boundaries. Although
significant effort has been devoted on different ways of capturing
and analyzing prior information, very little thought has been devoted
on the way of combining image information with prior information.
This paper focuses on a more natural way of incorporating the
prior information in the level set framework. For proof of concept
the method is applied on hippocampus segmentation in T1-MR
images. Hippocampus segmentation is a very challenging task, due
to the multivariate surrounding region and the missing boundary
with the neighboring amygdala, whose intensities are identical. The
proposed method, mimics the human segmentation way and thus
shows enhancements in the segmentation accuracy.
Abstract: This research is aimed to compare the percentages of correct classification of Empirical Bayes method (EB) to Classical method when data are constructed as near normal, short-tailed and long-tailed symmetric, short-tailed and long-tailed asymmetric. The study is performed using conjugate prior, normal distribution with known mean and unknown variance. The estimated hyper-parameters obtained from EB method are replaced in the posterior predictive probability and used to predict new observations. Data are generated, consisting of training set and test set with the sample sizes 100, 200 and 500 for the binary classification. The results showed that EB method exhibited an improved performance over Classical method in all situations under study.