Decision Trees

Idea in short

A decision tree is a visual representation of decisions and their qualitative and quantitative outcomes. Since the methodology structure is in the form of a tree, it is called a decision tree.

Contemporary management challenges are influenced, not only by isolated decisions, but also a sequence of decisions. Moreover, today’s decisions will set the stage for future decisions. Furthermore, uncertainty further accentuates the situation as our current learnings and experiences determine future decisions. Hence, we could represent issues are a tree of decisions.

Decision trees – a decision-making framework – enables transcending superficial, one-cause, one-effect relationships. They also help discover and describe things in the context of multiple influences. The Decision Tree Framework is a robust decision scaling approach that adopts a visual representation that helps thoroughly understand a project’s vulnerabilities. Thus, decision trees can help appraise management choices, risks, objectives, and monetary gains associated with a decision.

Origin

The history depicts decision trees as a powerful tool for interpreting results with an intuitive tree-like display that enhances understanding and disseminates results. The origins of decision trees trace back to the era of the development of written records. The Greek philosopher Porphyry devised a form of the decision tree, Porphyrian tree, which is the oldest known classification tree that dates back to the 3^rd century C.E.

Around the 1960s, the concept of decision trees as a decision-making tool developed. The earliest use of decision trees in management decision-making appears in articles “How to Use Decision Trees in Capital Investments,” and “Decision Trees for Decision Making” by John F. Magee, in the July-August and September-October issues of the Harvard Business Review published in 1964.

:{In the 20th:{ century, computational decision trees also emerged simultaneously as the developing fields of }:Artificial Intelligence:{ and }:Statistical Analysis:{. The computational origins of decision trees are models of biological and cognitive processes. }:Frank P. Ramsey:{, in 1931, was the first to develop a methodology of decision analysis. This was based on probability and utility. Modern versions of decision trees }:trace their roots:{ to the research done by }::{Dr. William Belson in the 1950s of nationwide audience surveys on behalf of the British Broadcasting Corporation.}:

The Model

A decision tree helps analyze options and determine steps in a process wherein each decision leads to another step in the decision-making process until management reaches a final decision. Decision trees allows considering both, favorable outcomes and potential consequences.

Characteristics

Decision trees earn their name from their resemblance to the trunks and branches of a tree. The main trunk is the decision to be made. Each main branch is one option that can be part of the decision-making process. The smaller branches represent potential outcomes. The decision point is called a node, like a fork in a tree branch. The associated terminologies with a decision tree are:

Root Node: Root node is from where the decision tree starts
Leaf Node: Leaf nodes are the final output node
Splitting: Splitting is the process of dividing the decision node into sub-nodes according to the given conditions
Sub Tree: Branch is a tree formed by splitting the tree
Pruning: Pruning is the removal of unwanted branches from the tree

Following is the description of symbols used in a decision tree:

Tools

Decision trees can be created by simply drawing them on a whiteboard or piece of paper. Today, however, there are several digital tools exist to create decision trees. These tools provide an excellent medium to gather feedback and suggestions from the team and real-time collaboration. The digital tools include add-ins for Spreadsheets, websites with templates or online drawing tools, and web applications.

Methodology

The process of making a decision tree is as follows:

Identify all the options to complete the project

Make a set of relevant variables or decision parameters and include all the existing options. These variables will act as nodes of the tree

Assess the relative importance of variables

A hierarchy needs to be established between the parameters based on evaluating the results each option will bring. The most important parameter will be the root node, and the rest of the nodes are assigned accordingly based on decreasing importance

Analyze the results

Analyze each of the results presented in the decision tree. By considering the limitations and objectives of the project, determine how acceptable the different outcomes are

Optimize decisions

This involves determining options that best fit the project. The decision tree is used to decide which options have the highest probability of success and the most significant benefits

A Decision Tree Problem

The problem statement involves purchasing either an apartment, office or warehouse. Following are the steps to follow to make a decision using the decision tree:

:{...

decision-trees