Decision Tree is a graphical representation for obtaining all feasible solutions to a problem or decision given certain conditions. It is a Supervised learning technique which can be used to solve both classification and regression problems, but it is most commonly used to solve classification problems. It is termed a decision tree because of its root node and tree-like structure.
Decision is a tree-structured classifier, with internal nodes representing dataset features, branches representing decision rules, and leaf nodes representing outcomes. The Decision Node and the Leaf Node are the two nodes in a Decision tree. Decision nodes are used to make any decision and have multiple branches, whereas Leaf nodes are the results of those decisions and have no further branches. The tests are made based on the characteristics of the given dataset. Where the need to predict an answer arises, the decision divides itself into subtrees to give an answer based on Yes/No.
Let’s assume there is a financial institution making a decision to determine whether to grant a loan request or not which should be based on the credit score of the loan applicants a result, the decision tree begins at the root node to answer this problem (Attribute Selection Measure by Age). Based on the corresponding labels, the root node splits into the next decision node ( Average Salary/Income) and one leaf node. The following decision node is divided into one decision node (Credit history with other borrowers) and one leaf node. The decision node eventually continues to splits until the final two leaf nodes (Accepted offers and Declined offer). Consider the diagram below:
It is recommended to balance the dataset prior to fitting with the decision tree in a situation where decision tree learners create biased trees if some classes dominate over the other classes.
Step 1: Start with the root node, which holds the entire dataset, for example Y.
Step 2: Using the Attribute Selection Measure, find the best attribute in the dataset (ASM).
Step 3: Subdivide the S into subsets that contain the best attribute’s possible values.
Step 4: Create the node of the decision tree that has the best attribute.
Step 5: Create additional decision trees in a recursive manner using the subsets of the dataset obtained in step 3. Continue this process until the nodes can no longer be classified and the final node is designated as a leaf node.
While using the Decision Tree, we make the following assumptions:
At first, the entire training set is considered the root.
- Categorized feature values are desired; if the values persist, they are changed to discrete values before the model is built.
- Records are recursively distributed based on attribute values.
- As a root node or an internal node, we employ a statistical method for sorting attributes.
- It is straightforward to comprehend since it follows the identical steps that a human would take while making a decision in real life.
2. It can be extremely helpful in resolving decision-making issues.
3. It is beneficial to consider all of the possible solutions to an issue.
4. In comparison to other methods, data cleansing is not required as much.
1. The decision tree is complicated since it has multiple tiers.
2. It might have an overfitting problem, which the Random Forest method can help with.
3. The decision tree’s computing complexity may increase as the number of class labels grows.
1. Identifying potential growth opportunities: One of the applications of decision trees is evaluating potential business growth opportunities based on past data. Historical sales data can be utilized in decision trees to help businesses expand and grow by allowing them to make significant adjustments in their strategy.
2. Finding prospective clients using demographic data
Another usage of decision trees is in the identification of potential clients using demographic data. They can assist in streamlining a marketing budget and making informed selections about the business’s target market. Without decision trees, the company may spend its marketing budget without considering a specific demographic, affecting overall sales.
3. Providing assistance in a variety of fields
They can assist in finding the best tactics to help a firm reach its objectives. For example, Financial Institution can also use decision trees to estimate the likelihood of a customer defaulting on a loan by generating predictive models based on the client’s previous data. Financial Institutions can utilize a decision tree support tool to evaluate a customer’s creditworthiness.
In business operations, decision trees can be used to plan logistics and strategic management. Engineering, education, law, business, healthcare, and finance are some of the other sectors where decision trees can be used.
references: datadriveninvestor.com, analyticsvidhya.com