The Role of Decision Tree in Data Mining
Branches of possibilities and outcomes
Weall make decisions in life depending upon the possibilities and possible outcomes. The track of all these possibilities and outcomes take the form of a tree metaphorically. The role of Decision Tree in Data Mining is tremendous. Read a brief account of this important current topic in data science.
What is decision Tree?
To decide on a process execution one goes through various possibilities of decisions, this leads to the formation of decision tree. The possibilities are “Yes(Y)”, “No(N)” or “either/or”. For example on a cross road there are two possibilities either left or right road. If left road taken then what will be the result and if right road is taken what will be the result. The decision making may change the outcomes. One can also predict the decision making process of that passenger if he has been following passenger for long through his decisioning analysis.
One can then make a visual map of all decisions taken by explicitly representing all moves. Visual map is a tree like representation of decision model
Decision Tree Rules
Decision tree generate rules. These are conditional statements generated by decision analysis that humans can understand and which are based in a database, to identify set of records. The decision rule is a simple IF and Then statement based on a condition and a prediction. For example if it has rained today then it will rain tomorrow(prediction) as it is a rainy season (condition).
Advantage of Decision Tree in Data Mining
The advantage of decision tree is that it depicts all possible conditions in a decision making, Predicts all possible outcomes and provides each path that leads to conclusion or making decision. This unique quality is useful in following areas:
- Businesses in making predictions
- Analyzing risk and risk mitigation
- Machine learning technology
- Artificial Intelligence
- Data Mining
- Text Mining
- Information Extraction.
These decision tree are useful in preparation procedures to supply administered wisdom in machine learning.
The decision tree are used in supervised machine learning models. It helps to predict the decisions by analyzing previous paths taken by data sets. Model is trained and tested on a set of test data that contains desired categorization and then applied to New set of data for predictions.
Decision tree is a plan or set of algorithms with root nodes, branch and leaf nodes. Each leaf nodes takes a class tag.
Each internal node characterizes examination of an attribute. Primary node is the tree root node. All decision tree have three nodes:
Decision Node: Shows a decision and represented by square or diamond shape.
Chance Node: Shows a chance or confusion in decision and represented by a circle.
End Node: Shows a result and represented by a triangle.
Decision tree is used in Data Mining, for text extraction, Information extraction. It is most powerful tool for data classification and prediction. Decision tree and knowledge graphs are tools that break the silos in data sets and allow more understanding and connections between data and help in making important business insights!
