Decision Trees
Introduction
A decision tree[1] is a prediction model used in various fields ranging from artificial intelligence to economics or law.[2] Given a set of data, diagrams of logical constructions are manufactured, very similar to rule-based prediction systems, which serve to represent and categorize a series of conditions that occur successively, for the resolution of a problem.
Construction of a decision tree
Contenido
Vamos a explicar cómo se construye un árbol de decisión. Para ello, vamos a hacer hincapié en varios aspectos.
Items
Decision trees are made up of nodes, vectors of numbers, arrows, and labels.
Concepts
When we deal with tree development we will frequently use these concepts:
Rules
In decision trees a series of rules must be met.
In decision trees, the decisions that are chosen are linear; as you select between several options, others are closed, which normally implies that there is no going back. In general it could be said that the rules follow a conditional form: Option 1->option 2->option 3->Final Result
These rules are usually implicit in the data set from which the decision tree is built.
Example
In this decision tree we can see how there are four possible final solutions. It complies with the rules described above (initial node, a single arrow per node and a single path to reach each final node) and has all the elements described above (nodes, number vectors, arrows and labels). The image represents a decision tree made up of two players, in it you can see the options that
The game has two players. The first decision must be made by player 1, who must decide between O1 and O2, at this point it will be player 2 who decides. If Player 1 has chosen O1, she will have to decide between A1 and R1; from 0 to two. With the appropriate existing induction methods, this decision tree could be solved without much complication.