Abstract: Many factors affect the success of Machine Learning
(ML) on a given task. The representation and quality of the instance
data is first and foremost. If there is much irrelevant and redundant
information present or noisy and unreliable data, then knowledge
discovery during the training phase is more difficult. It is well known
that data preparation and filtering steps take considerable amount of
processing time in ML problems. Data pre-processing includes data
cleaning, normalization, transformation, feature extraction and
selection, etc. The product of data pre-processing is the final training
set. It would be nice if a single sequence of data pre-processing
algorithms had the best performance for each data set but this is not
happened. Thus, we present the most well know algorithms for each
step of data pre-processing so that one achieves the best performance
for their data set.
Abstract: Rule Discovery is an important technique for mining
knowledge from large databases. Use of objective measures for
discovering interesting rules leads to another data mining problem,
although of reduced complexity. Data mining researchers have
studied subjective measures of interestingness to reduce the volume
of discovered rules to ultimately improve the overall efficiency of
KDD process.
In this paper we study novelty of the discovered rules as a
subjective measure of interestingness. We propose a hybrid approach
based on both objective and subjective measures to quantify novelty
of the discovered rules in terms of their deviations from the known
rules (knowledge). We analyze the types of deviation that can arise
between two rules and categorize the discovered rules according to
the user specified threshold. We implement the proposed framework
and experiment with some public datasets. The experimental results
are promising.
Abstract: A clustering is process to identify a homogeneous
groups of object called as cluster. Clustering is one interesting topic
on data mining. A group or class behaves similarly characteristics.
This paper discusses a robust clustering process for data images with
two reduction dimension approaches; i.e. the two dimensional
principal component analysis (2DPCA) and principal component
analysis (PCA). A standard approach to overcome this problem is
dimension reduction, which transforms a high-dimensional data into
a lower-dimensional space with limited loss of information. One of
the most common forms of dimensionality reduction is the principal
components analysis (PCA). The 2DPCA is often called a variant of
principal component (PCA), the image matrices were directly treated
as 2D matrices; they do not need to be transformed into a vector so
that the covariance matrix of image can be constructed directly using
the original image matrices. The decomposed classical covariance
matrix is very sensitive to outlying observations. The objective of
paper is to compare the performance of robust minimizing vector
variance (MVV) in the two dimensional projection PCA (2DPCA)
and the PCA for clustering on an arbitrary data image when outliers
are hiden in the data set. The simulation aspects of robustness and
the illustration of clustering images are discussed in the end of
paper
Abstract: Need for an appropriate system of evaluating students-
educational developments is a key problem to achieve the predefined
educational goals. Intensity of the related papers in the last years; that
tries to proof or disproof the necessity and adequacy of the students
assessment; is the corroborator of this matter. Some of these studies
tried to increase the precision of determining question weights in
scientific examinations. But in all of them there has been an attempt
to adjust the initial question weights while the accuracy and precision
of those initial question weights are still under question. Thus In
order to increase the precision of the assessment process of students-
educational development, the present study tries to propose a new
method for determining the initial question weights by considering
the factors of questions like: difficulty, importance and complexity;
and implementing a combined method of PROMETHEE and fuzzy
analytic network process using a data mining approach to improve
the model-s inputs. The result of the implemented case study proves
the development of performance and precision of the proposed
model.
Abstract: For the past one decade, biclustering has become popular data mining technique not only in the field of biological data analysis but also in other applications like text mining, market data analysis with high-dimensional two-way datasets. Biclustering clusters both rows and columns of a dataset simultaneously, as opposed to traditional clustering which clusters either rows or columns of a dataset. It retrieves subgroups of objects that are similar in one subgroup of variables and different in the remaining variables. Firefly Algorithm (FA) is a recently-proposed metaheuristic inspired by the collective behavior of fireflies. This paper provides a preliminary assessment of discrete version of FA (DFA) while coping with the task of mining coherent and large volume bicluster from web usage dataset. The experiments were conducted on two web usage datasets from public dataset repository whereby the performance of FA was compared with that exhibited by other population-based metaheuristic called binary Particle Swarm Optimization (PSO). The results achieved demonstrate the usefulness of DFA while tackling the biclustering problem.
Abstract: This work deals with aspects of support vector learning for large-scale data mining tasks. Based on a decomposition algorithm that can be run in serial and parallel mode we introduce a data transformation that allows for the usage of an expensive generalized kernel without additional costs. In order to speed up the decomposition algorithm we analyze the problem of working set selection for large data sets and analyze the influence of the working set sizes onto the scalability of the parallel decomposition scheme. Our modifications and settings lead to improvement of support vector learning performance and thus allow using extensive parameter search methods to optimize classification accuracy.
Abstract: With the hardware technology advancing, the cost of
storing is decreasing. Thus there is an urgent need for new techniques
and tools that can intelligently and automatically assist us in
transferring this data into useful knowledge. Different techniques of
data mining are developed which are helpful for handling these large
size databases [7]. Data mining is also finding its role in the field of
biotechnology. Pedigree means the associated ancestry of a crop
variety. Genetic diversity is the variation in the genetic composition
of individuals within or among species. Genetic diversity depends
upon the pedigree information of the varieties. Parents at lower
hierarchic levels have more weightage for predicting genetic
diversity as compared to the upper hierarchic levels. The weightage
decreases as the level increases. For crossbreeding, the two varieties
should be more and more genetically diverse so as to incorporate the
useful characters of the two varieties in the newly developed variety.
This paper discusses the searching and analyzing of different possible
pairs of varieties selected on the basis of morphological characters,
Climatic conditions and Nutrients so as to obtain the most optimal
pair that can produce the required crossbreed variety. An algorithm
was developed to determine the genetic diversity between the
selected wheat varieties. Cluster analysis technique is used for
retrieving the results.
Abstract: Outlier detection in streaming data is very challenging because streaming data cannot be scanned multiple times and also new concepts may keep evolving. Irrelevant attributes can be termed as noisy attributes and such attributes further magnify the challenge of working with data streams. In this paper, we propose an unsupervised outlier detection scheme for streaming data. This scheme is based on clustering as clustering is an unsupervised data mining task and it does not require labeled data, both density based and partitioning clustering are combined for outlier detection. In this scheme partitioning clustering is also used to assign weights to attributes depending upon their respective relevance and weights are adaptive. Weighted attributes are helpful to reduce or remove the effect of noisy attributes. Keeping in view the challenges of streaming data, the proposed scheme is incremental and adaptive to concept evolution. Experimental results on synthetic and real world data sets show that our proposed approach outperforms other existing approach (CORM) in terms of outlier detection rate, false alarm rate, and increasing percentages of outliers.
Abstract: Biclustering is a very useful data mining technique for
identifying patterns where different genes are co-related based on a
subset of conditions in gene expression analysis. Association rules
mining is an efficient approach to achieve biclustering as in
BIMODULE algorithm but it is sensitive to the value given to its
input parameters and the discretization procedure used in the
preprocessing step, also when noise is present, classical association
rules miners discover multiple small fragments of the true bicluster,
but miss the true bicluster itself. This paper formally presents a
generalized noise tolerant bicluster model, termed as μBicluster. An
iterative algorithm termed as BIDENS based on the proposed model
is introduced that can discover a set of k possibly overlapping
biclusters simultaneously. Our model uses a more flexible method to
partition the dimensions to preserve meaningful and significant
biclusters. The proposed algorithm allows discovering biclusters that
hard to be discovered by BIMODULE. Experimental study on yeast,
human gene expression data and several artificial datasets shows that
our algorithm offers substantial improvements over several
previously proposed biclustering algorithms.
Abstract: Clustering is the process of subdividing an input data set into a desired number of subgroups so that members of the same subgroup are similar and members of different subgroups have diverse properties. Many heuristic algorithms have been applied to the clustering problem, which is known to be NP Hard. Genetic algorithms have been used in a wide variety of fields to perform clustering, however, the technique normally has a long running time in terms of input set size. This paper proposes an efficient genetic algorithm for clustering on very large data sets, especially on image data sets. The genetic algorithm uses the most time efficient techniques along with preprocessing of the input data set. We test our algorithm on both artificial and real image data sets, both of which are of large size. The experimental results show that our algorithm outperforms the k-means algorithm in terms of running time as well as the quality of the clustering.
Abstract: Data mining can be called as a technique to extract
information from data. It is the process of obtaining hidden
information and then turning it into qualified knowledge by statistical
and artificial intelligence technique. One of its application areas is
medical area to form decision support systems for diagnosis just by
inventing meaningful information from given medical data. In this
study a decision support system for diagnosis of illness that make use
of data mining and three different artificial intelligence classifier
algorithms namely Multilayer Perceptron, Naive Bayes Classifier and
J.48. Pima Indian dataset of UCI Machine Learning Repository was
used. This dataset includes urinary and blood test results of 768
patients. These test results consist of 8 different feature vectors.
Obtained classifying results were compared with the previous studies.
The suggestions for future studies were presented.
Abstract: Network security attacks are the violation of
information security policy that received much attention to the
computational intelligence society in the last decades. Data mining
has become a very useful technique for detecting network intrusions
by extracting useful knowledge from large number of network data
or logs. Naïve Bayesian classifier is one of the most popular data
mining algorithm for classification, which provides an optimal way
to predict the class of an unknown example. It has been tested that
one set of probability derived from data is not good enough to have
good classification rate. In this paper, we proposed a new learning
algorithm for mining network logs to detect network intrusions
through naïve Bayesian classifier, which first clusters the network
logs into several groups based on similarity of logs, and then
calculates the prior and conditional probabilities for each group of
logs. For classifying a new log, the algorithm checks in which cluster
the log belongs and then use that cluster-s probability set to classify
the new log. We tested the performance of our proposed algorithm by
employing KDD99 benchmark network intrusion detection dataset,
and the experimental results proved that it improves detection rates
as well as reduces false positives for different types of network
intrusions.
Abstract: In an era of knowledge explosion, the growth of data
increases rapidly day by day. Since data storage is a limited resource,
how to reduce the data space in the process becomes a challenge issue.
Data compression provides a good solution which can lower the
required space. Data mining has many useful applications in recent
years because it can help users discover interesting knowledge in large
databases. However, existing compression algorithms are not
appropriate for data mining. In [1, 2], two different approaches were
proposed to compress databases and then perform the data mining
process. However, they all lack the ability to decompress the data to
their original state and improve the data mining performance. In this
research a new approach called Mining Merged Transactions with the
Quantification Table (M2TQT) was proposed to solve these problems.
M2TQT uses the relationship of transactions to merge related
transactions and builds a quantification table to prune the candidate
itemsets which are impossible to become frequent in order to improve
the performance of mining association rules. The experiments show
that M2TQT performs better than existing approaches.
Abstract: There are several approaches in trying to solve the
Quantitative 1Structure-Activity Relationship (QSAR) problem.
These approaches are based either on statistical methods or on
predictive data mining. Among the statistical methods, one should
consider regression analysis, pattern recognition (such as cluster
analysis, factor analysis and principal components analysis) or partial
least squares. Predictive data mining techniques use either neural
networks, or genetic programming, or neuro-fuzzy knowledge. These
approaches have a low explanatory capability or non at all. This
paper attempts to establish a new approach in solving QSAR
problems using descriptive data mining. This way, the relationship
between the chemical properties and the activity of a substance
would be comprehensibly modeled.
Abstract: Mining frequent tree patterns have many useful
applications in XML mining, bioinformatics, network routing, etc.
Most of the frequent subtree mining algorithms (i.e. FREQT,
TreeMiner and CMTreeMiner) use anti-monotone property in the
phase of candidate subtree generation. However, none of these
algorithms have verified the correctness of this property in tree
structured data. In this research it is shown that anti-monotonicity
does not generally hold, when using weighed support in tree pattern
discovery. As a result, tree mining algorithms that are based on this
property would probably miss some of the valid frequent subtree
patterns in a collection of trees. In this paper, we investigate the
correctness of anti-monotone property for the problem of weighted
frequent subtree mining. In addition we propose W3-Miner, a new
algorithm for full extraction of frequent subtrees. The experimental
results confirm that W3-Miner finds some frequent subtrees that the
previously proposed algorithms are not able to discover.
Abstract: The myoelectric signal (MES) is one of the Biosignals
utilized in helping humans to control equipments. Recent approaches
in MES classification to control prosthetic devices employing pattern
recognition techniques revealed two problems, first, the classification
performance of the system starts degrading when the number of
motion classes to be classified increases, second, in order to solve the
first problem, additional complicated methods were utilized which
increase the computational cost of a multifunction myoelectric
control system. In an effort to solve these problems and to achieve a
feasible design for real time implementation with high overall
accuracy, this paper presents a new method for feature extraction in
MES recognition systems. The method works by extracting features
using Wavelet Packet Transform (WPT) applied on the MES from
multiple channels, and then employs Fuzzy c-means (FCM)
algorithm to generate a measure that judges on features suitability for
classification. Finally, Principle Component Analysis (PCA) is
utilized to reduce the size of the data before computing the
classification accuracy with a multilayer perceptron neural network.
The proposed system produces powerful classification results (99%
accuracy) by using only a small portion of the original feature set.
Abstract: In this paper, a data mining model to SMEs for detecting financial and operational risk indicators by data mining is presenting. The identification of the risk factors by clarifying the relationship between the variables defines the discovery of knowledge from the financial and operational variables. Automatic and estimation oriented information discovery process coincides the definition of data mining. During the formation of model; an easy to understand, easy to interpret and easy to apply utilitarian model that is far from the requirement of theoretical background is targeted by the discovery of the implicit relationships between the data and the identification of effect level of every factor. In addition, this paper is based on a project which was funded by The Scientific and Technological Research Council of Turkey (TUBITAK).
Abstract: The purpose of this research aims to discover the
knowledge for analysis student motivation behavior on e-Learning
based on Data Mining Techniques, in case of the Information
Technology for Communication and Learning Course at Suan
Sunandha Rajabhat University. The data mining techniques was
applied in this research including association rules, classification
techniques. The results showed that using data mining technique can
indicate the important variables that influence the student motivation
behavior on e-Learning.
Abstract: Data clustering is an important data exploration
technique with many applications in data mining. The k-means
algorithm is well known for its efficiency in clustering large data
sets. However, this algorithm is suitable for spherical shaped clusters
of similar sizes and densities. The quality of the resulting clusters
decreases when the data set contains spherical shaped with large
variance in sizes. In this paper, we introduce a competent procedure
to overcome this problem. The proposed method is based on shifting
the center of the large cluster toward the small cluster, and recomputing
the membership of small cluster points, the experimental
results reveal that the proposed algorithm produces satisfactory
results.
Abstract: Evidence-based medicine is a new direction in modern healthcare. Its task is to prevent, diagnose and medicate diseases using medical evidence. Medical data about a large patient population is analyzed to perform healthcare management and medical research. In order to obtain the best evidence for a given disease, external clinical expertise as well as internal clinical experience must be available to the healthcare practitioners at right time and in the right manner. External evidence-based knowledge can not be applied directly to the patient without adjusting it to the patient-s health condition. We propose a data warehouse based approach as a suitable solution for the integration of external evidence-based data sources into the existing clinical information system and data mining techniques for finding appropriate therapy for a given patient and a given disease. Through integration of data warehousing, OLAP and data mining techniques in the healthcare area, an easy to use decision support platform, which supports decision making process of care givers and clinical managers, is built. We present three case studies, which show, that a clinical data warehouse that facilitates evidence-based medicine is a reliable, powerful and user-friendly platform for strategic decision making, which has a great relevance for the practice and acceptance of evidence-based medicine.