Decision Tree Algorithm: A Survey
PunamKhandar
1
, Chetana Thaokar
2
and Abhijeet Thakare
3
1-3
Shri Ramdeobaba College of Engineering and Management, Nagpur, India
Email: khandarp@rknec.edu, thaokarcb@rknec.edu, thakarear@rknec.edu
Abstract—Decision tree is one of the popular techniques in the research community to be used
to develop decision making approaches. Over the period of almost four decades, a wide variety
of applications of decision trees for providing good quality solutions in the different areas of
life including technological pitfall, health hazards , social & geographical hitches, etc have been
tried by several authors. The kind of work to be stated includes Object recognition using binary
DT, Face detection, Diagnosing failures in large Internet sites, Prognostic decision making in
breast cancers, Speech emotion recognition, Flood probability mapping, Automated Detection
of Deceptive Language, Knowledge Discovery on Climate and Air Pollution and Fault
Detection. Some pieces of works have even put commendable effort for development of DT and
considerable evolution in the technique is observed. Some noteworthy works are effective
decision tree design technique with the help of priori statistics, overcoming the shortcoming of
ID3’s using a new DT algorithm that united the ID3 and Association Function (AF) as well as
an algorithm that combines principle of Taylor Formula with information entropy solution of
ID3, alternating decision tree (ADTree) technique, etc. Here we are making a sincere effort to
present some important applications of DT discussed by different authors through their
valuable work as well as the variety of evolutionary aspects of DT put forward by them for the
entire research community.
Index Terms— Decision tree, algorithms, design techniques, knowledge discovery.
I. INTRODUCTION
Decision trees are a powerful and popular mechanism for describing data. The rules derived from the decision
trees can be easily understood and even implemented directly as a database query for retrieval purposes [8].It is a
predictive model and used as a tool for decision support. Decision tree is important tool in data mining
approaches. It’s a tree-like model of decisions that recursively partitions the given space to subspaces that leads
to decision making. Such algorithms contain conditional control statements in which each internal node performs
a test on an attribute, each branch represents the result of the test, and each terminal node represents a class label.
Thus the classification rule is confirmed as the paths from root to leaf. Decision trees are essential supervised
paradigm of machine learning model, also popular one. It primly covers the areas like data mining, machine
learning, and pattern recognition. Originally developed in decision theory and statistics, DTs were enhanced by
researchers in Data science too for exploring large and complex bodies of data that identifies usefulpatterns.
The study contains 7 sections. Sect. 1 is introduction. Sect. 2 states various DT algorithms. Sect. 3 discusses the
evolution of DT over the decades. Sect. 4 is about different applications of DT in various domains. Sect. 5
focuses on important methodologies. Sect. 6 discusses the comparative study results. Sect. 7 is about conclusion
of the survey.
Grenze ID: 01.GIJET.8.2.47_1
© Grenze Scientific Society, 2022
Grenze International Journal of Engineering and Technology, June Issue