Abstract: Feature selection has recently been the subject of intensive research in data mining, specially for datasets with a large number of attributes. Recent work has shown that feature selection can have a positive effect on the performance of machine learning algorithms. The success of many learning algorithms in their attempts to construct models of data, hinges on the reliable identification of a small set of highly predictive attributes. The inclusion of irrelevant, redundant and noisy attributes in the model building process phase can result in poor predictive performance and increased computation. In this paper, a novel feature search procedure that utilizes the Ant Colony Optimization (ACO) is presented. The ACO is a metaheuristic inspired by the behavior of real ants in their search for the shortest paths to food sources. It looks for optimal solutions by considering both local heuristics and previous knowledge. When applied to two different classification problems, the proposed algorithm achieved very promising results.
Abstract: This paper presents a new strategy of identification
and classification of pathological voices using the hybrid method
based on wavelet transform and neural networks. After speech
acquisition from a patient, the speech signal is analysed in order to
extract the acoustic parameters such as the pitch, the formants, Jitter,
and shimmer. Obtained results will be compared to those normal and
standard values thanks to a programmable database. Sounds are
collected from normal people and patients, and then classified into
two different categories. Speech data base is consists of several
pathological and normal voices collected from the national hospital
“Rabta-Tunis". Speech processing algorithm is conducted in a
supervised mode for discrimination of normal and pathology voices
and then for classification between neural and vocal pathologies
(Parkinson, Alzheimer, laryngeal, dyslexia...). Several simulation
results will be presented in function of the disease and will be
compared with the clinical diagnosis in order to have an objective
evaluation of the developed tool.
Abstract: Segmentation, filtering out of measurement errors and
identification of breakpoints are integral parts of any analysis of
microarray data for the detection of copy number variation (CNV).
Existing algorithms designed for these tasks have had some successes
in the past, but they tend to be O(N2) in either computation time or
memory requirement, or both, and the rapid advance of microarray
resolution has practically rendered such algorithms useless. Here we
propose an algorithm, SAD, that is much faster and much less thirsty
for memory – O(N) in both computation time and memory requirement
-- and offers higher accuracy. The two key ingredients of SAD are the
fundamental assumption in statistics that measurement errors are
normally distributed and the mathematical relation that the product of
two Gaussians is another Gaussian (function). We have produced a
computer program for analyzing CNV based on SAD. In addition to
being fast and small it offers two important features: quantitative
statistics for predictions and, with only two user-decided parameters,
ease of use. Its speed shows little dependence on genomic profile.
Running on an average modern computer, it completes CNV analyses
for a 262 thousand-probe array in ~1 second and a 1.8 million-probe
array in 9 seconds
Abstract: Music Information Retrieval (MIR) and modern data mining techniques are applied to identify style markers in midi music for stylometric analysis and author attribution. Over 100 attributes are extracted from a library of 2830 songs then mined using supervised learning data mining techniques. Two attributes are identified that provide high informational gain. These attributes are then used as style markers to predict authorship. Using these style markers the authors are able to correctly distinguish songs written by the Beatles from those that were not with a precision and accuracy of over 98 per cent. The identification of these style markers as well as the architecture for this research provides a foundation for future research in musical stylometry.
Abstract: This paper presents a Neural Network (NN) identification of icing parameters in an A340 aircraft and a reconfiguration technique to keep the A/C performance close to the performance prior to icing. Five aircraft parameters are assumed to be considerably affected by icing. The off-line training for identifying the clear and iced dynamics is based on the Levenberg-Marquard Backpropagation algorithm. The icing parameters are located in the system matrix. The physical locations of the icing are assumed at the right and left wings. The reconfiguration is based on the technique known as the control mixer approach or pseudo inverse technique. This technique generates the new control input vector such that the A/C dynamics is not much affected by icing. In the simulations, the longitudinal and lateral dynamics of an Airbus A340 aircraft model are considered, and the stability derivatives affected by icing are identified. The simulation results show the successful NN identification of the icing parameters and the reconfigured flight dynamics having the similar performance before the icing. In other words, the destabilizing icing affect is compensated.
Abstract: The perfect operation of common Active Filters is depended on accuracy of identification system distortion. Also, using a suitable method in current injection and reactive power compensation, leads to increased filter performance. Due to this fact, this paper presents a method based on predictive current control theory in shunt active filter applications. The harmonics of the load current is identified by using o–d–q reference frame on load current and eliminating the DC part of d–q components. Then, the rest of these components deliver to predictive current controller as a Threephase reference current by using Park inverse transformation. System is modeled in discreet time domain. The proposed method has been tested using MATLAB model for a nonlinear load (with Total Harmonic Distortion=20%). The simulation results indicate that the proposed filter leads to flowing a sinusoidal current (THD=0.15%) through the source. In addition, the results show that the filter tracks the reference current accurately.
Abstract: In this paper, a new approach is introduced to solve
Blasius equation using parameter identification of a nonlinear
function which is used as approximation function. Bees Algorithm
(BA) is applied in order to find the adjustable parameters of
approximation function regarding minimizing a fitness function
including these parameters (i.e. adjustable parameters). These
parameters are determined how the approximation function has to
satisfy the boundary conditions. In order to demonstrate the
presented method, the obtained results are compared with another
numerical method. Present method can be easily extended to solve a
wide range of problems.
Abstract: Writer identification is one of the areas in pattern
recognition that attract many researchers to work in, particularly in
forensic and biometric application, where the writing style can be
used as biometric features for authenticating an identity. The
challenging task in writer identification is the extraction of unique
features, in which the individualistic of such handwriting styles
can be adopted into bio-inspired generalized global shape for
writer identification. In this paper, the feasibility of generalized
global shape concept of complimentary binding in Artificial
Immune System (AIS) for writer identification is explored. An
experiment based on the proposed framework has been conducted
to proof the validity and feasibility of the proposed approach for
off-line writer identification.
Abstract: Globalization, supported by information and
communication technologies, changes the rules of competitiveness
and increases the significance of information, knowledge and
network cooperation. In line with this trend, the need for efficient
trust-building tools has emerged. The absence of trust building
mechanisms and strategies was identified within several studies.
Through trust development, participation on e-business network and
usage of network services will increase and provide to SMEs new
economic benefits. This work is focused on effective trust building
strategies development for electronic business network platforms.
Based on trust building mechanism identification, the questionnairebased
analysis of its significance and minimum level of requirements
was conducted. In the paper, we are confirming the trust dependency
on e-Skills which play crucial role in higher level of trust into the
more sophisticated and complex trust building ICT solutions.
Abstract: Stable bacterial polymorphism on a single limiting resource may appear if between the evolved strains metabolic interactions take place that allow the exchange of essential nutrients [8]. Towards an attempt to predict the possible outcome of longrunning evolution experiments, a network based on the metabolic capabilities of homogeneous populations of every single gene knockout strain (nodes) of the bacterium E. coli is reconstructed. Potential metabolic interactions (edges) are allowed only between strains of different metabolic capabilities. Bacterial communities are determined by finding cliques in this network. Growth of the emerged hypothetical bacterial communities is simulated by extending the metabolic flux balance analysis model of Varma et al [2] to embody heterogeneous cell population growth in a mutual environment. Results from aerobic growth on 10 different carbon sources are presented. The upper bounds of the diversity that can emerge from single-cloned populations of E. coli such as the number of strains that appears to metabolically differ from most strains (highly connected nodes), the maximum clique size as well as the number of all the possible communities are determined. Certain single gene deletions are identified to consistently participate in our hypothetical bacterial communities under most environmental conditions implying a pattern of growth-condition- invariant strains with similar metabolic effects. Moreover, evaluation of all the hypothetical bacterial communities under growth on pyruvate reveals heterogeneous populations that can exhibit superior growth performance when compared to the performance of the homogeneous wild-type population.
Abstract: In this paper we propose a method for recognition of
adult video based on support vector machine (SVM). Different kernel
features are proposed to classify adult videos. SVM has an advantage
that it is insensitive to the relative number of training example in
positive (adult video) and negative (non adult video) classes. This
advantage is illustrated by comparing performance between different
SVM kernels for the identification of adult video.
Abstract: The current research paper is an implementation of
Eigen Faces and Karhunen-Loeve Algorithm for face recognition.
The designed program works in a manner where a unique
identification number is given to each face under trial. These faces
are kept in a database from where any particular face can be matched
and found out of the available test faces. The Karhunen –Loeve
Algorithm has been implemented to find out the appropriate right
face (with same features) with respect to given input image as test
data image having unique identification number. The procedure
involves usage of Eigen faces for the recognition of faces.
Abstract: This paper presents a new version of the SVM mixture algorithm initially proposed by Kwok for classification and regression problems. For both cases, a slight modification of the mixture model leads to a standard SVM training problem, to the existence of an exact solution and allows the direct use of well known decomposition and working set selection algorithms. Only the regression case is considered in this paper but classification has been addressed in a very similar way. This method has been successfully applied to engine pollutants emission modeling.
Abstract: Two approaches for model development of a smart acoustic box are suggested in this paper: the finite element (FE) approach and the subspace identification. Both approaches result in a state-space model, which can be used for obtaining the frequency responses and for the controller design. In order to validate the developed FE model and to perform the subspace identification, an experimental set-up with the acoustic box and dSPACE system was used. Experimentally obtained frequency responses show good agreement with the frequency responses obtained from the FE model and from the identified model.
Abstract: Phytases are acid phosphatase enzymes, which
efficiently cleave phosphate moieties from phytic acid, thereby
generating myo-inositol and inorganic phosphate. Thirty four
isolates of endophytic fungi to produce of phytases were isolated
from leaf, stem and root fragments of soybean. Screening of 34
isolates of endophytic fungi identified the phytases produced by
Rhizoctonia sp. and Fusarium verticillioides . The phytase
production were the best induced by phytic acid and rice bran
compared the others inducer in submerged fermentation medium
used. The phytase produced by both Rhizoctonia sp. and F.
verticillioides have pH optimum at 4.0 and 5.0 respectively. The
characterization of phytase from Fusarium verticillioides showed that
temperature optimum was 500C and stability until 600C, the pH
optimum 5.0 and pH stability was 2.5 – 6.0, and substrate specificity
were rice bran>soybean meal>corn> coconut cake, respectively.
Abstract: In this paper, we propose a novel algorithm for
delineating the endocardial wall from a human heart ultrasound scan.
We assume that the gray levels in the ultrasound images are
independent and identically distributed random variables with
different Rician Inverse Gaussian (RiIG) distributions. Both synthetic
and real clinical data will be used for testing the algorithm. Algorithm
performance will be evaluated using the expert radiologist evaluation
of a soft copy of an ultrasound scan during the scanning process and
secondly, doctor’s conclusion after going through a printed copy of
the same scan. Successful implementation of this algorithm should
make it possible to differentiate normal from abnormal soft tissue and
help disease identification, what stage the disease is in and how best
to treat the patient. We hope that an automated system that uses this
algorithm will be useful in public hospitals especially in Third World
countries where problems such as shortage of skilled radiologists and
shortage of ultrasound machines are common. These public hospitals
are usually the first and last stop for most patients in these countries.
Abstract: In recent years, Radio Frequency Identification (RFID)
is followed with interest by many researches, especially for the
purpose of indoor positioning as the innate properties of RFID are
profitable for achieving it. A lot of algorithms or schemes are proposed
to be used in the RFID-based positioning system, but most of them are
lack of environmental consideration and it induces inaccuracy of
application. In this research, a lot of algorithms and schemes of RFID
indoor positioning are discussed to see whether effective or not on
application, and some rules are summarized for achieving accurate
positioning. On the other hand, a new term “Noise Factor" is involved
to describe the signal loss between the target and the obstacle. As a
result, experimental data can be obtained but not only simulation; and
the performance of the positioning system can be expressed
substantially.
Abstract: The applications on numbers are across-the-board that there is much scope for study. The chic of writing numbers is diverse and comes in a variety of form, size and fonts. Identification of Indian languages scripts is challenging problems. In Optical Character Recognition [OCR], machine printed or handwritten characters/numerals are recognized. There are plentiful approaches that deal with problem of detection of numerals/character depending on the sort of feature extracted and different way of extracting them. This paper proposes a recognition scheme for handwritten Hindi (devnagiri) numerals; most admired one in Indian subcontinent our work focused on a technique in feature extraction i.e. Local-based approach, a method using 16-segment display concept, which is extracted from halftoned images & Binary images of isolated numerals. These feature vectors are fed to neural classifier model that has been trained to recognize a Hindi numeral. The archetype of system has been tested on varieties of image of numerals. Experimentation result shows that recognition rate of halftoned images is 98 % compared to binary images (95%).
Abstract: This paper presents an exact pruning algorithm with
adaptive pruning interval for general dynamic neural networks
(GDNN). GDNNs are artificial neural networks with internal dynamics.
All layers have feedback connections with time delays to the
same and to all other layers. The structure of the plant is unknown, so
the identification process is started with a larger network architecture
than necessary. During parameter optimization with the Levenberg-
Marquardt (LM) algorithm irrelevant weights of the dynamic neural
network are deleted in order to find a model for the plant as
simple as possible. The weights to be pruned are found by direct
evaluation of the training data within a sliding time window. The
influence of pruning on the identification system depends on the
network architecture at pruning time and the selected weight to be
deleted. As the architecture of the model is changed drastically during
the identification and pruning process, it is suggested to adapt the
pruning interval online. Two system identification examples show
the architecture selection ability of the proposed pruning approach.
Abstract: In this paper we use the property of co-occurrence
matrix in finding parallel lines in binary pictures for fingerprint
identification. In our proposed algorithm, we reduce the noise by
filtering the fingerprint images and then transfer the fingerprint
images to binary images using a proper threshold. Next, we divide
the binary images into some regions having parallel lines in the same
direction. The lines in each region have a specific angle that can be
used for comparison. This method is simple, performs the
comparison step quickly and has a good resistance in the presence of
the noise.