International Journal of Distributed Sensor Networks
Classication
Decision tree
ID3
C4.5
SLIQ
SPRINT
CART AID
CHAID
KNN
WKPDS
ENNS
EENNS
EEENNS
Bayesian
network
Bayesian
multinets
SVM
GSVM
FSVM
TWSVMs
VaR-SVM
RSVM
One-
Bayesian
dependence
Naïve Bayes
Selective
naïve Bayes
Bayes
Seminaïve
k-dependence
Bayesian
F : e research structure of classication.
(i) A decision tree is a ow-chart-like tree structure,
where each internal node is denoted by rectangles and
leaf nodes are denoted by ovals. All internal nodes
have two or more child nodes. All internal nodes
contain splits, which test the value of an expression
of the attributes. Arcs from an internal node to its
children are labeled with distinct outcomes of the test.
Each leaf node has a class label associated with it.
Iterative Dichotomiser or ID is a simple decision
tree learning algorithm []. C. algorithm is an
improved version of ID; it uses gain ratio as splitting
criteria []. e dierence between ID and C.
algorithm is that ID uses binary splits, whereas C.
algorithm uses multiway splits. SLIQ (Supervised
Learning In Quest) is capable of handling large data
sets with ease and lesser time complexity [, ],
SPRINT (Scalable Parallelizable Induction of Deci-
sion Tree algorithm) is also fast and highly scalable,
and there is no storage constraint on larger data sets
in SPRINT []. Other improvement researches are
nished [, ]. Classication and Regression Trees
(CART) is a nonparametric decision tree algorithm.
It produces either classication or regression trees,
basedonwhethertheresponsevariableiscategor-
ical or continuous. CHAID (chi-squared automatic
interaction detector) and the improvement researcher
[] focus on dividing a data set into exclusive and
exhaustive segments that dier with respect to the
response variable.
(ii) e KNN (K-Nearest Neighbor) algorithm is intro-
duced by the Nearest Neighbor algorithm which is
designed to nd the nearest point of the observed
object.emainideaoftheKNNalgorithmistond
the K-nearest points [].erearealotofdierent
improvements for the traditional KNN algorithm,
such as the Wavelet Based K-Nearest Neighbor Partial
Distance Search (WKPDS) algorithm [], Equal-
Average Nearest Neighbor Search (ENNS) algorithm
[], Equal-Average Equal-Norm Nearest Neighbor
code word Search (EENNS) algorithm [], the
Equal-Average Equal-Variance Equal-Norm Nearest
Neighbor Search (EEENNS) algorithm [], and
other improvements [].
(iii) Bayesian networks are directed acyclic graphs whose
nodes represent random variables in the Bayesian
sense. Edges represent conditional dependencies;
nodes which are not connected represent vari-
ables which are conditionally independent of each
other. Based on Bayesian networks, these classiers
have many strengths, like model interpretability and
accommodation to complex data and classication
problem settings []. e research includes na
¨
ıve
Bayes [, ], selective na
¨
ıve Bayes [], semina
¨
ıve
Bayes [], one-dependence Bayesian classiers [,
], K-dependence Bayesian classiers [], Bayesian
network-augmented na
¨
ıve Bayes [], unrestricted
Bayesian classiers [], and Bayesian multinets [].
(iv) Support Vector Machines algorithm is supervised
learning model with associated learning algorithms
that analyze data and recognize patterns, which is
based on statistical learning theory. SVM produces
a binary classier, the so-called optimal separating
hyperplanes, through an extremely nonlinear map-
ping of the input vectors into the high-dimensional
feature space []. SVM is widely used in text
classication [, ], marketing, pattern recogni-
tion, and medical diagnosis []. A lot of further
research is done, GSVM (granular support vector
machines) [–], FSVM (fuzzy support vector
machines) [–], TWSVMs (twin support vector
machines) [–], VaR-SVM (value-at-risk support