AbstractThis study established concept elements
based on granular computing theory and the isomorphic
relation between rated scales in formal concept analysis
(FCA) and constructed the correlation of the concept elements.
A concept granule was constructed by studying
the mapping relation between concept elements. The common
polymerization and extension forms of the concept
granule were given. We studied the condition in which the
granular structure in a conceptual system...This study established concept elements
based on granular computing theory and the isomorphic
relation between rated scales in formal concept analysis
(FCA) and constructed the correlation of the concept elements.
A concept granule was constructed by studying
the mapping relation between concept elements. The common
polymerization and extension forms of the concept
granule were given. We studied the condition in which the
granular structure in a conceptual system is purified, as
well as the formation mechanism and generalized cohesiveness
of concept granules. An algorithm for classification
rule discovery algorithm based on concept granule
structure-the GRD algorithm-was created. According
to experimental results, the proposed GRD algorithm has
higher classification accuracy, simpler rule set, and better
generalization than traditional algorithms for classification
rule discovery. The formal representation based
on conceptual elements shows that a knowledge representation
model that is complete in terms of semantic
description can be built.Read MoreRead Less
AbstractCollaborative filtering (CF) is the most successful
and widely utilized recommendation technology.
CF-based recommenders perform well in terms of accuracy,
but they lack the capability to find fresh and novel
items. To improve the novel recommendation of userbased
CF, the definition of a novel item was established,
and an appropriate strategy of novel recommendation was
determined. First, a novel item containing the three aspects
of likability, unknown, and...Collaborative filtering (CF) is the most successful
and widely utilized recommendation technology.
CF-based recommenders perform well in terms of accuracy,
but they lack the capability to find fresh and novel
items. To improve the novel recommendation of userbased
CF, the definition of a novel item was established,
and an appropriate strategy of novel recommendation was
determined. First, a novel item containing the three aspects
of likability, unknown, and dissimilarity was defined
based on the meaning of the term novel. Second, metrics
that measure the novelty of the recommendation system
were designed based on the timeliness of novelty. Finally,
for the comparison of different strategies of novel recommendation,
unknown and dissimilarity were integrated into
the process and output of traditional algorithms. Results
show that the novelty of the recommendation system is
significantly improved when unknown and dissimilarity are
integrated into the recommendation results of the traditional
algorithm to recalculate the novelty of the item and
set the accuracy threshold. Output integration strategy
can improve the novelty of the recommended results and
can be utilized for any algorithm.Read MoreRead Less
AbstractABSTRACT: Classification of imbalanced data has been
recognized as a crucial problem in machine learning and
data mining.In an imbalanced dataset, minority class instances
are likely to be misclassified. When the synthetic
minority over-sampling technique (SMOTE) is applied
in imbalanced dataset classification, the same sampling
rate is set for all samples of the minority class in
the process of synthesizing new samples, this scenario
involves blindness. To overcome...ABSTRACT: Classification of imbalanced data has been
recognized as a crucial problem in machine learning and
data mining.In an imbalanced dataset, minority class instances
are likely to be misclassified. When the synthetic
minority over-sampling technique (SMOTE) is applied
in imbalanced dataset classification, the same sampling
rate is set for all samples of the minority class in
the process of synthesizing new samples, this scenario
involves blindness. To overcome this problem, an improved
SMOTE algorithm based on genetic algorithm (GA),
namely, GASMOTE was proposed. First, GASMOTE set
different sampling rates for different minority class
samples. A combination of the sampling rates corresponded
to an individual in the population. Second, the
selection, crossover, and mutation operators of GA were
iteratively applied to the population to obtain the best
combination of sampling rates when the stopping criteria
were met. Lastly, the best combination of sampling rates
was used in SMOTE to synthetize new samples. Experimental
results on 10 typical imbalanced datasets show
that GASMOTE increases the F-measure value by 5.9%
and the G-mean value by 1.6% compared with the SMOTE
algorithm. Meanwhile, GASMOTE increases the F-measure
value by 3.7% and the G-mean value by 2.3% compared
with the borderline-SMOTE algorithm. GASMOTE
can be utilized as a new over-sampling technique to address
the problem of imbalanced dataset classification.
Read MoreRead Less
AbstractFuzzy Petri nets (FPNs) are an ideal modeling
tool for knowledge-based systems, which are based
on fuzzy production rules. FPNs are widely used in knowledge
representation and reasoning, assessment, fault
diagnosis, exception handling, and other fields, but they
have the defects of single membership degree. To solve
this problem, intuitionistic fuzzy Petri nets (IFPNs) were
presented for knowledge representation and reasoning.
First, the IFPN model was constructed for...Fuzzy Petri nets (FPNs) are an ideal modeling
tool for knowledge-based systems, which are based
on fuzzy production rules. FPNs are widely used in knowledge
representation and reasoning, assessment, fault
diagnosis, exception handling, and other fields, but they
have the defects of single membership degree. To solve
this problem, intuitionistic fuzzy Petri nets (IFPNs) were
presented for knowledge representation and reasoning.
First, the IFPN model was constructed for knowledge representation
by combining intuitionistic fuzzy sets theory
with Petri nets theory. Second, an algorithm based on
IFPN was proposed, and the matrix operation was introduced
into the reasoning process to make full use of the
parallel computing capability of Petri nets. Finally, an example
was illustrated to prove the feasibility and advantages
of the proposed IFPN model and reasoning algorithm.
Moreover, the reasoning result was analyzed and
discussed. Compared with FPN, IFPN can describe three
states, namely, the support state, the opposite state, and
the neutral state. Thus, IFPN can overcome the single
membership degree of FPN and describe the reasoning
result more comprehensively and precisely than FPN.
Moreover, IFPN is an effective extension and development
of FPN and will become a promising method for
knowledge representation and reasoning.Read MoreRead Less
AbstractThe aim of the work is to increase the reliability
of the AES cipher by means of development and
application of redundant codes of a polynomial residue
number system (PRNS) that are able to correct the errors
caused by failures.The known methods of counteracting
failures do not take into account the specificities
of the AES cipher, which leads to significant hardware
costs. The problem can be solved...The aim of the work is to increase the reliability
of the AES cipher by means of development and
application of redundant codes of a polynomial residue
number system (PRNS) that are able to correct the errors
caused by failures.The known methods of counteracting
failures do not take into account the specificities
of the AES cipher, which leads to significant hardware
costs. The problem can be solved through the use of
error-correcting codes of a polynomial residue number
system. However, it is impossible to use the known methods
of search and correction of errors by the PRNS codes
in the AES cipher. Therefore, the development of a method,
the use of which will give the AES cipher the robustness
property with respectto failures through the use of the
PRNS codes, is a topical problem.Read MoreRead Less
AbstractPlagiarism detection is a sensitive field of
research which has gained lot of interest in the past few
years. Although plagiarism detection systems are developed
to check text in a variety of languages, they perform
better when they are dedicated to check a specific language
as they take into account the specificity of the
language which leads to better quality results. Query optimization
and document reduction constitute...Plagiarism detection is a sensitive field of
research which has gained lot of interest in the past few
years. Although plagiarism detection systems are developed
to check text in a variety of languages, they perform
better when they are dedicated to check a specific language
as they take into account the specificity of the
language which leads to better quality results. Query optimization
and document reduction constitute two major
processing modules which play a major role in optimizing
the response time and the results quality of these systems
and hence determine their efficiency and effectiveness.
This paper proposes an analysis of approaches,
an architecture, and a system for detecting plagiarism in
Arabic documents. This analysis is particularly focused
on the methods and techniques used to detect plagiarism.
The proposed web-based architecture exhibits the
major processing modules of a plagiarism detection system
which are articulated into four layers inside a processing
component. The architecture has been used to
develop a plagiarism detection system for the Arabic language
proposing a set of functions to the user for checking
a text and analyzing the results through a well-designed
graphical user interface.Read MoreRead Less
AbstractWeibo is an extensively used social network
tool in China and has become a popular platform for disaster
information management. This popular
microblogging service offers massive firsthand information
regarding the state and emotions of victims in a disaster
situation. Identifying negative sentiment messages
from the large-scale and noisy Weibo stream is a fundamental
and challenging undertaking. Therefore, based on
the characteristics of negative Weibo messages concerning
disaster events, a...Weibo is an extensively used social network
tool in China and has become a popular platform for disaster
information management. This popular
microblogging service offers massive firsthand information
regarding the state and emotions of victims in a disaster
situation. Identifying negative sentiment messages
from the large-scale and noisy Weibo stream is a fundamental
and challenging undertaking. Therefore, based on
the characteristics of negative Weibo messages concerning
disaster events, a novel feature selection algorithm
called combined frequent pattern (FP)-growth and mutual
information theory (CFM) algorithm, was proposed to improve
the traditional machine learning approaches in this
study. The CFM algorithm mined two FPs via FP-tree,
and the mutual information between two frequent items
was calculated to determine the most frequent and tight
features for negative-sentiment Weibo messages detection.
After that, the experimental analysis was conducted
to test the proposed novel feature selection algorithm and
to explore a suitable sentiment classifier for disaster-related
Weibo messages. The analysis employed actual
disaster-related Weibo message data set, which included
2,913 negative messages and 2,913 un-negative messages.
Results demonstrate that the CFM algorithm performs
well in the feature selection process. In particular,
this algorithm exhibits the best performance in the support
vector machine classifier with 89.34% accuracy.
Therefore, the CFM algorithm is an efficient feature selection
algorithm for negative-message classification in a
post-disaster situation. This algorithm also offers a novel
method to reduce the feature dimension in other text classification
areas.Read MoreRead Less