Abstract: Clustering algorithms are attractive for the task of class identification in spatial databases. However, the application to large spatial databases rises the following requirements for clustering algorithms: minimal requirements of domain knowledge to determine the input parameters, discovery of clusters with arbitrary shape and good efficiency on large databases. The well-known clustering algorithms offer no solution to the combination of these requirements. In this paper, a density based clustering algorithm (DCBRD) is presented, relying on a knowledge acquired from the data by dividing the data space into overlapped regions. The proposed algorithm discovers arbitrary shaped clusters, requires no input parameters and uses the same definitions of DBSCAN algorithm. We performed an experimental evaluation of the effectiveness and efficiency of it, and compared this results with that of DBSCAN. The results of our experiments demonstrate that the proposed algorithm is significantly efficient in discovering clusters of arbitrary shape and size.
Abstract: Wireless sensor networks include small nodes which
have sensing ability; calculation and connection extend themselves
everywhere soon. Such networks have source limitation on
connection, calculation and energy consumption. So, since the nodes
have limited energy in sensor networks, the optimized energy
consumption in these networks is of more importance and has created
many challenges. The previous works have shown that by organizing
the network nodes in a number of clusters, the energy consumption
could be reduced considerably. So the lifetime of the network would
be increased. In this paper, we used the Queen-bee algorithm to
create energy efficient clusters in wireless sensor networks. The
Queen-bee (QB) is similar to nature in that the queen-bee plays a
major role in reproduction process. The QB is simulated with J-sim
simulator. The results of the simulation showed that the clustering by
the QB algorithm decreases the energy consumption with regard to
the other existing algorithms and increases the lifetime of the
network.
Abstract: In this paper we present a new approach to detecting a
flaw in T.O.F.D (Time Of Flight Diffraction) type ultrasonic image
based on texture features. Texture is one of the most important
features used in recognizing patterns in an image. The paper
describes texture features based on 2D Gabor functions, i.e.,
Gaussian shaped band-pass filters, with dyadic treatment of the radial
spatial frequency range and multiple orientations, which represent an
appropriate choice for tasks requiring simultaneous measurement in
both space and frequency domains. The most relevant features are
used as input data on a Fuzzy c-mean clustering classifier. The
classes that exist are only two: 'defects' or 'no defects'. The proposed
approach is tested on the T.O.F.D image achieved at the laboratory
and on the industrial field.
Abstract: Cluster analysis is the name given to a diverse collection of techniques that can be used to classify objects (e.g. individuals, quadrats, species etc). While Kohonen's Self-Organizing Feature Map (SOFM) or Self-Organizing Map (SOM) networks have been successfully applied as a classification tool to various problem domains, including speech recognition, image data compression, image or character recognition, robot control and medical diagnosis, its potential as a robust substitute for clustering analysis remains relatively unresearched. SOM networks combine competitive learning with dimensionality reduction by smoothing the clusters with respect to an a priori grid and provide a powerful tool for data visualization. In this paper, SOM is used for creating a toroidal mapping of two-dimensional lattice to perform cluster analysis on results of a chemical analysis of wines produced in the same region in Italy but derived from three different cultivators, referred to as the “wine recognition data" located in the University of California-Irvine database. The results are encouraging and it is believed that SOM would make an appealing and powerful decision-support system tool for clustering tasks and for data visualization.
Abstract: With deep development of software reuse, componentrelated
technologies have been widely applied in the development of
large-scale complex applications. Component identification (CI) is
one of the primary research problems in software reuse, by analyzing
domain business models to get a set of business components with high
reuse value and good reuse performance to support effective reuse.
Based on the concept and classification of CI, its technical stack is
briefly discussed from four views, i.e., form of input business models,
identification goals, identification strategies, and identification
process. Then various CI methods presented in literatures are
classified into four types, i.e., domain analysis based methods,
cohesion-coupling based clustering methods, CRUD matrix based
methods, and other methods, with the comparisons between these
methods for their advantages and disadvantages. Additionally, some
insufficiencies of study on CI are discussed, and the causes are
explained subsequently. Finally, it is concluded with some
significantly promising tendency about research on this problem.
Abstract: The development of aid's systems for the medical
diagnosis is not easy thing because of presence of inhomogeneities in
the MRI, the variability of the data from a sequence to the other as
well as of other different source distortions that accentuate this
difficulty. A new automatic, contextual, adaptive and robust
segmentation procedure by MRI brain tissue classification is
described in this article. A first phase consists in estimating the
density of probability of the data by the Parzen-Rozenblatt method.
The classification procedure is completely automatic and doesn't
make any assumptions nor on the clusters number nor on the
prototypes of these clusters since these last are detected in an
automatic manner by an operator of mathematical morphology called
skeleton by influence zones detection (SKIZ). The problem of
initialization of the prototypes as well as their number is transformed
in an optimization problem; in more the procedure is adaptive since it
takes in consideration the contextual information presents in every
voxel by an adaptive and robust non parametric model by the
Markov fields (MF). The number of bad classifications is reduced by
the use of the criteria of MPM minimization (Maximum Posterior
Marginal).
Abstract: Although there have been many researches in cluster
analysis to consider on feature weights, little effort is made on sample
weights. Recently, Yu et al. (2011) considered a probability
distribution over a data set to represent its sample weights and then
proposed sample-weighted clustering algorithms. In this paper, we
give a sample-weighted version of generalized fuzzy clustering
regularization (GFCR), called the sample-weighted GFCR
(SW-GFCR). Some experiments are considered. These experimental
results and comparisons demonstrate that the proposed SW-GFCR is
more effective than the most clustering algorithms.
Abstract: A neuron can emit spikes in an irregular time basis and by averaging over a certain time window one would ignore a lot of information. It is known that in the context of fast information processing there is no sufficient time to sample an average firing rate of the spiking neurons. The present work shows that the spiking neurons are capable of computing the radial basis functions by storing the relevant information in the neurons' delays. One of the fundamental findings of the this research also is that when using overlapping receptive fields to encode the data patterns it increases the network-s clustering capacity. The clustering algorithm that is discussed here is interesting from computer science and neuroscience point of view as well as from a perspective.
Abstract: Advances in clinical medical imaging have brought about the routine production of vast numbers of medical images that need to be analyzed. As a result an enormous amount of computer vision research effort has been targeted at achieving automated medical image analysis. Computed Tomography (CT) is highly accurate for diagnosing liver tumors. This study aimed to evaluate the potential role of the wavelet and the neural network in the differential diagnosis of liver tumors in CT images. The tumors considered in this study are hepatocellular carcinoma, cholangio carcinoma, hemangeoma and hepatoadenoma. Each suspicious tumor region was automatically extracted from the CT abdominal images and the textural information obtained was used to train the Probabilistic Neural Network (PNN) to classify the tumors. Results obtained were evaluated with the help of radiologists. The system differentiates the tumor with relatively high accuracy and is therefore clinically useful.
Abstract: Eukaryotic protein-coding genes are interrupted by spliceosomal introns, which are removed from the RNA transcripts before translation into a protein. The exon-intron structures of different eukaryotic species are quite different from each other, and the evolution of such structures raises many questions. We try to address some of these questions using statistical analysis of whole genomes. We go through all the protein-coding genes in a genome and study correlations between the net length of all the exons in a gene, the number of the exons, and the average length of an exon. We also take average values of these features for each chromosome and study correlations between those averages on the chromosomal level. Our data show universal features of exon-intron structures common to animals, plants, and protists (specifically, Arabidopsis thaliana, Caenorhabditis elegans, Drosophila melanogaster, Cryptococcus neoformans, Homo sapiens, Mus musculus, Oryza sativa, and Plasmodium falciparum). We have verified linear correlation between the number of exons in a gene and the length of a protein coded by the gene, while the protein length increases in proportion to the number of exons. On the other hand, the average length of an exon always decreases with the number of exons. Finally, chromosome clustering based on average chromosome properties and parameters of linear regression between the number of exons in a gene and the net length of those exons demonstrates that these average chromosome properties are genome-specific features.
Abstract: Clustering algorithms help to understand the hidden
information present in datasets. A dataset may contain intrinsic and
nested clusters, the detection of which is of utmost importance. This
paper presents a Distributed Grid-based Density Clustering algorithm
capable of identifying arbitrary shaped embedded clusters as well as
multi-density clusters over large spatial datasets. For handling
massive datasets, we implemented our method using a 'sharednothing'
architecture where multiple computers are interconnected
over a network. Experimental results are reported to establish the
superiority of the technique in terms of scale-up, speedup as well as
cluster quality.
Abstract: Optimization is often a critical issue for most system
design problems. Evolutionary Algorithms are population-based,
stochastic search techniques, widely used as efficient global
optimizers. However, finding optimal solution to complex high
dimensional, multimodal problems often require highly
computationally expensive function evaluations and hence are
practically prohibitive. The Dynamic Approximate Fitness based
Hybrid EA (DAFHEA) model presented in our earlier work [14]
reduced computation time by controlled use of meta-models to
partially replace the actual function evaluation by approximate
function evaluation. However, the underlying assumption in
DAFHEA is that the training samples for the meta-model are
generated from a single uniform model. Situations like model
formation involving variable input dimensions and noisy data
certainly can not be covered by this assumption. In this paper we
present an enhanced version of DAFHEA that incorporates a
multiple-model based learning approach for the SVM approximator.
DAFHEA-II (the enhanced version of the DAFHEA framework) also
overcomes the high computational expense involved with additional
clustering requirements of the original DAFHEA framework. The
proposed framework has been tested on several benchmark functions
and the empirical results illustrate the advantages of the proposed
technique.
Abstract: This paper develops a quality estimation method with
the application of fuzzy hierarchical clustering. Quality estimation is
essential to quality control and quality improvement as a precise
estimation can promote a right decision-making in order to help
better quality control. Normally the quality of finished products in
manufacturing system can be differentiated by quality standards. In
the real life situation, the collected data may be vague which is not
easy to be classified and they are usually represented in term of fuzzy
number. To estimate the quality of product presented by fuzzy
number is not easy. In this research, the trapezoidal fuzzy numbers
are collected in manufacturing process and classify the collected data
into different clusters so as to get the estimation. Since normal
hierarchical clustering methods can only be applied for real numbers,
fuzzy hierarchical clustering is selected to handle this problem based
on quality standards.
Abstract: Students in high education are presented with new terms and concepts in nearly every lecture they attend. Many of them prefer Web-based self-tests for evaluation of their concepts understanding since they can use those tests independently of tutors- working hours and thus avoid the necessity of being in a particular place at a particular time. There is a large number of multiple-choice tests in almost every subject designed to contribute to higher level learning or discover misconceptions. Every single test provides immediate feedback to a student about the outcome of that test. In some cases a supporting system displays an overall score in case a test is taken several times by a student. What we still find missing is how to secure delivering of personalized feedback to a user while taking into consideration the user-s progress. The present work is motivated to throw some light on that question.
Abstract: This paper deals with the application for contentbased
image retrieval to extract color feature from natural images
stored in the image database by segmenting the image through
clustering. We employ a class of nonparametric techniques in which
the data points are regarded as samples from an unknown probability
density. Explicit computation of the density is avoided by using the
mean shift procedure, a robust clustering technique, which does not
require prior knowledge of the number of clusters, and does not
constrain the shape of the clusters. A non-parametric technique for
the recovery of significant image features is presented and
segmentation module is developed using the mean shift algorithm to
segment each image. In these algorithms, the only user set parameter
is the resolution of the analysis and either gray level or color images
are accepted as inputs. Extensive experimental results illustrate
excellent performance.
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: The Chiu-s method which generates a Takagi-Sugeno Fuzzy Inference System (FIS) is a method of fuzzy rules extraction. The rules output is a linear function of inputs. In addition, these rules are not explicit for the expert. In this paper, we develop a method which generates Mamdani FIS, where the rules output is fuzzy. The method proceeds in two steps: first, it uses the subtractive clustering principle to estimate both the number of clusters and the initial locations of a cluster centers. Each obtained cluster corresponds to a Mamdani fuzzy rule. Then, it optimizes the fuzzy model parameters by applying a genetic algorithm. This method is illustrated on a traffic network management application. We suggest also a Mamdani fuzzy rules generation method, where the expert wants to classify the output variables in some fuzzy predefined classes.
Abstract: Clustering techniques have been used by many intelligent software agents to group similar access patterns of the Web users into high level themes which express users intentions and interests. However, such techniques have been mostly focusing on one salient feature of the Web document visited by the user, namely the extracted keywords. The major aim of these techniques is to come up with an optimal threshold for the number of keywords needed to produce more focused themes. In this paper we focus on both keyword and similarity thresholds to generate themes with concentrated themes, and hence build a more sound model of the user behavior. The purpose of this paper is two fold: use distance based clustering methods to recognize overall themes from the Proxy log file, and suggest an efficient cut off levels for the keyword and similarity thresholds which tend to produce more optimal clusters with better focus and efficient size.
Abstract: The rapid urbanization of cities has a bane in the form
road accidents that cause extensive damage to life and limbs. A
number of location based factors are enablers of road accidents in the
city. The speed of travel of vehicles is non-uniform among locations
within a city. In this study, the perception of vehicle users is captured
on a 10-point rating scale regarding the degree of variation in speed
of travel at chosen locations in the city. The average rating is used to
cluster locations using fuzzy c-means clustering and classify them as
low, moderate and high speed of travel locations. The high speed of
travel locations can be classified proactively to ensure that accidents
do not occur due to the speeding of vehicles at such locations. The
advantage of fuzzy c-means clustering is that a location may be a
part of more than one cluster to a varying degree and this gives a
better picture about the location with respect to the characteristic
(speed of travel) being studied.
Abstract: In this paper, we propose an energy efficient cluster
based communication protocol for wireless sensor network. Our
protocol considers both the residual energy of sensor nodes and the
distance of each node from the BS when selecting cluster-head. This
protocol can successfully prolong the network-s lifetime by 1)
reducing the total energy dissipation on the network and 2) evenly
distributing energy consumption over all sensor nodes. In this
protocol, the nodes with more energy and less distance from the BS
are probable to be selected as cluster-head. Simulation results with
MATLAB show that proposed protocol could increase the lifetime of
network more than 94% for first node die (FND), and more than 6%
for the half of the nodes alive (HNA) factor as compared with
conventional protocols.