rnew icon6Grab Deal : Flat 30% off on live classes + 2 free self-paced courses! - SCHEDULE CALL rnew icon7


Classification by Decision Tree Induction in Data Mining


To learn decision trees using training tuples annotated with classes, we use a technique called induction. Each internal node (non-leaf node) of a decision tree in data mining represents a test on an attribute, each branch represents a possible result of that test, and each leaf node (or terminal node) stores a class label. The node at the top of a tree is called the root node. Understanding decision tree data mining begins with understanding data science; you can get an insight into the same through our Data Science Training.

Decision Tree Induction

In data mining, the supervised learning technique of decision tree is utilized for classification and regression. There's a tree there that aids with selecting choices. Tree-like models for classification or deterioration can be generated using the decision tree. While creating the decision tree, the data set is partitioned into smaller subgroups. As a result, we get a decision tree with leaf nodes upon completion.

At the very least, a decision node will have two forks. The leaf nodes represent classification or a final verdict. Leaf nodes are the topmost decision nodes in a tree and are connected to the best predictor, also known as the root node. Decision trees are flexible enough to handle both numerical and categorical information.

Decision Tree in Data Mining

Depicts the structure of a typical data mining decision tree. It stands for the idea of buying a computer and can foretell whether or not an AllElectronics consumer will make that purchase.

Ovals represent leaf nodes, while rectangles represent internal nodes. Not all decision tree algorithms can generate nonbinary trees; some can only generate binary trees (in which each internal node branches to precisely two other nodes).

When asked, "How are decision trees used for categorizing objects?" The attribute values of a given tuple (X) are compared to those in the decision tree to determine whether or not the tuple belongs to a specific class. The data is followed up from the root via a series of intermediate nodes until it reaches a leaf node that stores the class prediction for that tuple. Classification rules may be generated using decision trees with little effort.

Why are decision trees used in so many classification systems? Decision tree classifier design is well-suited to discovery through exploration since it does not need prior knowledge of the domain or the establishment of parameters. High-dimensional data is no problem for decision trees. The tree structure they use to display learned information is natural and straightforward to understand for most people. Decision tree induction's learning and categorization processes are quick and easy.

In most cases, classification decisions made using decision trees are highly accurate. However, the quality of the available data may influence the outcome. For categorization purposes, several fields have turned to decision tree induction algorithms, including those in medicine, industry, economics, astronomy, and molecular biology.

Several popular rule induction tools are based on decision trees. You can also learn the six stages of data science processing to grasp the above topic better.

Algorithm For Decision Tree

Algorithm: Generate a decision tree. Generate a decision tree from the training tuples of data partition D.


1) Data partition, D, which is a set of training tuples and their associated class labels; 

2) Attribute list, the set of candidate attributes; 

3) Attribute selection method, a procedure to determine the splitting criterion that “best” partitions the data tuples into individual classes. This criterion consists of a splitting attribute and, possibly, a split point or splitting subset.

Output: A decision tree.


  •  (Step 1) Create a node N;

  •  (Step 2) If tuples in D are all of the same class, C then 

  •  (Step 3) Return N as a leaf node labeled with the class C;

  •  (Step 4) If attribute list is empty then 

  •  (Step 5) Return N as a leaf node labeled with the majority class in D; // majority voting

  •  (Step 6) Apply Attribute selection method(D, attribute list) to find the “best” splitting criterion;

  •  (Step 7) Label node N with splitting criterion;

  •  (Step 8) If splitting attribute is discrete-valued and multiway splits allowed, then // not restricted to  binary trees

  •  (Step 9) Attribute list ← attribute list − splitting attribute; // remove splitting attribute

  •  (Step 10) For each outcome j of splitting criterion // partition the tuples and grow subtrees for each partition

  •  (Step 11) Let Dj be the set of data tuples in D satisfying outcome j; // a partition

  •  (Step 12) If Dj is empty then

  •  (Step 13) Attach a leaf labeled with the majority class in D to node N;

  •  (Step 14) Else attach the node returned by Generate decision tree(Dj , attribute list) to node N; end for  

  • (Step 15) Return N;

4) There are three inputs to the algorithm:

The first one is dimension (D), a list of attributes, and a technique for selecting those attributes (Attribute selection method). D is shorthand for "data partition," which is the official name for this concept. At first, it is the full collection of training tuples and their corresponding class labels. The set of properties that characterize the tuples is the parameter attribute list. The attribute selection technique This method uses a metric for selecting attributes, like the Gini index or information gain. In most cases, the attribute selection measure determines whether or not the tree contains only binary nodes. The Gini index is one such attribute selection measure that requires the output tree to be binary. While certain factors limit branching, such as knowledge acquisition, others do not (i.e., two or more branches to be grown from a node).

5) The training tuples in D are initially represented by a single node in the tree, denoted by the letter N. (step 1).

6) When node N is a leaf and is assigned a class based on whether or not the tuples in D are all of the same class, we say that D is a class-based tree (steps 2 and 3). It's important to realize that 4 and 5 are dead ends. The algorithm's last section explains each of the possible exit situations.

7) In this case, the algorithm uses the Attribute selection approach to choose the criteria for the split. By identifying the "optimal" way to partition the tuples in D into individual classes, realize the splitting criteria indicate which attribute should be tested at node N. (step 6). With respect to the test results, the splitting criterion also specifies which branches to develop from node N. In particular, the splitting criterion denotes the splitting attribute and, optionally, a split-point or splitting subset. The splitting criteria are established to produce "pure" partitions at each branch. Pure partitions contain only tuples from the same category. In other words, we want the partitions that arise from splitting the tuples in D following the mutually exclusive results of the splitting criterion to be as clean as feasible.

cta10 icon

Data Science Training

  • Personalized Free Consultation
  • Access to Our Learning Management System
  • Access to Our Course Curriculum
  • Be a Part of Our Free Demo Class

 8) The splitting criterion is a test at attributed is indicated by the label (step 7). For each possible result of the splitting criterion, node N sprouts a new branch. Tuples in D are separated as needed (step 1 indicates, there are three distinct outcomes available if we define A as the dividing characteristic. According to the data in the training set, the value of A can take on v different forms: a1, a2,..., av.

  • If A has a finite number of values, then the test results at node N will map directly onto those values. A new branch is generated and labeled with the value aj for each value aj of A. The subset of class-labeled tuples in D with the value aj of A is denoted by the partition Dj. Because all tuples within a particular partition have the same value for A, A is no longer relevant to the partitioning process. As a result, it has been taken off the list of attributes (steps 8 to 9). 
  • If A is continuous, then there are two possible test results at node N: either A is less than or greater than the split point. When the Attribute selection method returns a split point as part of the splitting criterion, use that point as the split point. (In practice, the split-point, a, is commonly selected as the midpoint of two known adjacent values of A and may not be a pre-existing value of A from the training data.) Based on these findings, two new branches are developed from N and given appropriate labels. D1 contains the class-labeled tuples in D for which A split point and D2 contains the remaining tuples.
  • The value of A is discrete. Hence a binary tree must be constructed (per the requirements of the attribute selection measure or method being used): At node N, we do an "A SA?" test. When Attribute selection is used as the splitting criterion, it returns SA as the subset of A that may be divided into smaller pieces. This list contains only some of the A values currently known to exist. The test at node N is considered to be true if and only if the provided tuple has a value aj of A and aj SA.Shows that N is the parent of two separate branches. D1 is the subset of class-labeled tuples in D that pass the test, and it corresponds to the left branch out of N, which is always labeled yes. D2 is the subset of class-labeled tuples from D that fail the test because they stem from the right branch of N, labeled no.


9) The algorithm uses the same process recursively to form a decision tree for the tuples at each resulting partition, Dj, of D (step 14). 

The recursive partitioning stops only when any of the following terminating conditions is true:

  • All tuples in partition D (represented at node N) belong to the same class (steps 2 and 3),
  • There are no remaining attributes on which the tuples may be further partitioned (step 4). In this case, majority voting is employed (step 5). This involves converting node N into a leaf and labeling it with the most common class in D. Alternatively, the class distribution of the node tuples may be stored.
  • There are no tuples for a given branch; that is, a partition Dj is empty (step 12). In this case, a leaf is created with the majority class in D (step 13).

10) The resulting decision tree is returned (step 15).

Advantages of Using Decision Trees:

There are few advantages of using decision tree induction in data mining. Let’s explore these advantages:

  • Information scaling is unnecessary when using a decision tree.
  • Data with missing values do not hinder the decision tree construction process either.
  • Automatic and straightforward to explain to both the technical team and the stakeholders, a decision tree model can do the heavy lifting for you.
  • Decision trees need less work in pre-processing data than other methods.
  • To use a decision tree, it is not necessary to normalize the information being analyzed.


Decision tree induction plays an essential role in data mining by providing valuable insights into the complex relationships between input variables and outcomes. It has several advantages over traditional statistical methods, including ease of use and scalability. It is suitable for diverse applications ranging from finance/insurance sectors, fraud detection, healthcare industry, customer segmentation, marketing campaigns, etc. You can also learn about neural network guides and python for data science if you are interested in further career prospects in data science. Check Our community page for data science community 

Course Links:

Data Science Courses Course Links
Data Engineering Certification Training - Using R Or Python https://www.janbasktraining.com/data-engineering
Artificial Intelligence Online Certification Training


Python Online Training & Certification Course


Machine Learning Online Certification Training Course https://www.janbasktraining.com/machine-learning
Tableau With Data Science Training & Certification https://www.janbasktraining.com/data-visualization-with-tableau

Trending Courses

Cyber Security icon

Cyber Security

  • Introduction to cybersecurity
  • Cryptography and Secure Communication 
  • Cloud Computing Architectural Framework
  • Security Architectures and Models
Cyber Security icon1

Upcoming Class

16 days 05 Jul 2024

QA icon


  • Introduction and Software Testing
  • Software Test Life Cycle
  • Automation Testing and API Testing
  • Selenium framework development using Testing
QA icon1

Upcoming Class

8 days 27 Jun 2024

Salesforce icon


  • Salesforce Configuration Introduction
  • Security & Automation Process
  • Sales & Service Cloud
  • Apex Programming, SOQL & SOSL
Salesforce icon1

Upcoming Class

7 days 26 Jun 2024

Business Analyst icon

Business Analyst

  • BA & Stakeholders Overview
  • BPMN, Requirement Elicitation
  • BA Tools & Design Documents
  • Enterprise Analysis, Agile & Scrum
Business Analyst icon1

Upcoming Class

9 days 28 Jun 2024

MS SQL Server icon

MS SQL Server

  • Introduction & Database Query
  • Programming, Indexes & System Functions
  • SSIS Package Development Procedures
  • SSRS Report Design
MS SQL Server icon1

Upcoming Class

9 days 28 Jun 2024

Data Science icon

Data Science

  • Data Science Introduction
  • Hadoop and Spark Overview
  • Python & Intro to R Programming
  • Machine Learning
Data Science icon1

Upcoming Class

2 days 21 Jun 2024

DevOps icon


  • Intro to DevOps
  • GIT and Maven
  • Jenkins & Ansible
  • Docker and Cloud Computing
DevOps icon1

Upcoming Class

5 days 24 Jun 2024

Hadoop icon


  • Architecture, HDFS & MapReduce
  • Unix Shell & Apache Pig Installation
  • HIVE Installation & User-Defined Functions
  • SQOOP & Hbase Installation
Hadoop icon1

Upcoming Class

2 days 21 Jun 2024

Python icon


  • Features of Python
  • Python Editors and IDEs
  • Data types and Variables
  • Python File Operation
Python icon1

Upcoming Class

17 days 06 Jul 2024

Artificial Intelligence icon

Artificial Intelligence

  • Components of AI
  • Categories of Machine Learning
  • Recurrent Neural Networks
  • Recurrent Neural Networks
Artificial Intelligence icon1

Upcoming Class

10 days 29 Jun 2024

Machine Learning icon

Machine Learning

  • Introduction to Machine Learning & Python
  • Machine Learning: Supervised Learning
  • Machine Learning: Unsupervised Learning
Machine Learning icon1

Upcoming Class

23 days 12 Jul 2024

 Tableau icon


  • Introduction to Tableau Desktop
  • Data Transformation Methods
  • Configuring tableau server
  • Integration with R & Hadoop
 Tableau icon1

Upcoming Class

2 days 21 Jun 2024