Decision tree is a diagram which tries to display the range of probable results and consequent decisions made after the first decision. Decision tree contains tree main parts that include branches, leaf nodes and root node. The root node is the starting point of the tree and both leaf nodes and root have questions or criteria to be responded. Branches are arrows linking nodes, indicating the flow from question to response. Every node characteristically contains two or more nodes extending from it. For instance, if the question in the initial node needs a “no” or “yes” answer, there will be a leaf node, however, if there will be a leaf node if the answer is “no” and another for if the response is “yes” (Dimick, 2011).
A decision tree contains three forms of nodes and two forms of branches. A decision node is a section where a choice has to be made. Decision branches are the branches extending from a decision node. An event node is a section where uncertainty is resolved, where event branch is a branch contained in the event set up and that extends from an event node. Finally, a terminal node is the decision tree endpoints. It appears at the end of the decision tree graph.
Benefits of Using Decision Tree
A decision tree can be employed as model for a chronological decision issues under uncertainty. A decision tree graphically describes the decision to be made, the events that might happen, and the results related with combinations of events and decisions. Probabilities are allocated to the events, and values are established for each outcome. Decision tree major goal is to establish the best decision in a given situation. It offers a number of benefits to the user as compared to other analysis techniques. By use of graphical aspect of decision tree, one can schematically represent possible outcomes, chance events, and decision alternatives. The visual technique assists with understanding complex decision dependencies and sequences. Decision tree offers a high level of efficiency. With it, one can easily express a multifaceted decision problem clearly. In addition, one can quickly modify a decision tree as novel information turns to be available. After a decision tree is set, one can employ it to contrast how changing values of input influence the decision alternatives (Olivas, 2007).
Decision tree is also revealing. With it, one can contrast competing alternatives in terms of probable values and risks. The term expected value combines relative anticipated uncertainties and payoffs into a single arithmetical value. By so doing, it reveals the overall qualities of competing decision options. Decision tree also contains complementary values whereby, one can employ it in conjunction with other tools of project management.
Application of Decision Tree in Real Situation
A decision tree is a tree in which every branch node stands for a choice between several alternatives, and every leaf node stands for a decision or classification. It can therefore be employed in a number of real life situations. For instance, a decision tree can be employed to assist a financial institution to decide on whether an individual should be given a loan or not. In this case, a decision tree is employed to evaluate the applicant income, the number of years the person has been employed, and criminal record among other factors and evaluate the chances that the applicant will be able to repay the applied loan. It is also employed in astronomy to filter noise from the Hubble to obtain a clear Space Telescope images. It has also been employed in star-galaxy classification, evaluating galaxy counts and in determining quasars in the Second Palomar Sky Survey. Decision tree has also been applied in medical practice and research. The recent uses of automatic decision tree induction can be established in gastroenterology, psychiatry, cardiology, and diagnosis for detecting mammography microcalcifications, to diagnose thyroid disorders, and to examine sudden infant death (SID) syndrome. Beside these, decision tree technique has been employed in many other fields that include physics, pharmacology, agriculture, and molecular biology among others.
Science of Probability
Probability is an estimation or measure of a likelihood of an occurrence of a phenomenon. It give a value between 0 and 1 in percentage that a certain even is likely to take place. When a chance tends toward 1, the event is highly likely to take place while when it tends toward zero it is less likely to happen. It is therefore evident that probability aids in decision making and it is very reliable and effective in doing so. This is because, probability analyzes all possible approaches that can be given to a certain problem, and gives the possible outcomes ranging from 0 to 1. It therefore makes it clear on the outcomes that every possible decision would result to and thus, making it easy for an individual to pick the best possible solution to the problem (Erev & Wallsten, 1993).