Problems need to be addressed as they become available. But often it happens that each subsequent decision depends on the decision of the previous one, and in such a situation it is especially important to systematize tasks and to predict the results of these or those actions a few steps forward. This will help you with the unique method of decision tree.
Method for building a decision tree
Like any tree, the decision tree consists of "branches" and "leaves". Of course, drawing skills are not useful here, since the decision tree is a graphical systematization of the decision-making process, which reflects alternative solutions and environmental conditions, as well as possible risks and gains for any combination of these alternatives. In other words, it is an effective method of automatic data analysis (current and alternative), notable for its visibility.
Application of the decision tree
The decision tree is a popular method, applied in the most diverse spheres of our life:
- tree of administrative decisions is effective in project management, management, as well as in the analysis of all possible risks;
- the decision tree method has been used successfully to control the quality of products in the industry;
- in medicine, the decision tree is used to diagnose diseases;
- the decision tree method is applicable even for analysis of the structure of amino acids (molecular biology).
How to build a decision tree?
1. As a rule, the decision tree is located from right to left and does not contain cyclic elements (a new leaf or branch can only be split).
2. We need to start by showing the structure of the problem in the "trunk" of the future decision tree (right).
3. Branches are alternative solutions that can theoretically be adopted in a given situation, as well as the possible consequences of adopting these alternative solutions. Branches originate from one point (source data), but "grow" until the final result is obtained. The number of branches does not indicate the quality of your tree at all. In some cases (if the tree is too "branched"), it is recommended that you even use the clipping of secondary branches.
Branches come in two forms:
- Dotted lines that connect the squares - possible solutions;
- solid lines connecting the circles of possible final results.
4. Nodes are key events, and the lines connecting nodes are the works for the implementation of the project. Square nodes are the places where a decision is made. Round nodes are the appearance of results. Since, when making decisions, we can not influence the appearance of the outcome, we need to calculate the probability of their appearance.
5. In addition, in the decision tree, you need to display all the information about the time of work, their cost, as well as the probability of making each decision;
6. After all decisions and expected results are indicated on the tree, the analysis and choice of the most profitable way is carried out.
One of the most common tree models is the three-layer model, when the initial question is the first layer of possible solutions,
When making a decision tree, it is necessary to realize that the number of variants of the development of the situation must be observable and have some time limitation. In addition, the effectiveness of the method depends on the quality of the information put into the scheme.
An important advantage is that the decision tree can be combined with expert methods at stages requiring expert evaluation of the result. This increases the quality of the analysis of the decision tree and contributes to the correct choice of strategy.