decision tree


Also found in: Dictionary, Thesaurus, Medical, Legal, Acronyms, Encyclopedia, Wikipedia.

Decision tree

Schematic way of representing alternative sequential decisions and the possible outcomes from these decisions.

Decision Tree

In risk analysis, a diagram of decisions and their potential consequences. It is used to help determine the most straightforward (and cheapest) way to arrive at a stated goal. It is represented by potential decisions (drawn as squares), branching off into different proximate consequences (drawn as circles), and potential end results (drawn as triangles).
Fig.32 Decision tree. The businessman has two options: to open a new factory to boost production capacity or not to open a new factory; and he has to consider two states of nature or events which can occur economic boom or recession. The businessman must assess the likelihood of each of these events occurring and, in this case, based on his knowledge and experience, he estimates that there is a one-in-two chance of a boom and a 0.5 probability of a recession. Finally, the businessman estimates the financial consequences as an £80,000 profit for the new factory if there is a boom, and a £30,000 loss if there is a recession.click for a larger image
Fig.32 Decision tree. The businessman has two options: to open a new factory to boost production capacity or not to open a new factory; and he has to consider two states of nature or events which can occur economic boom or recession. The businessman must assess the likelihood of each of these events occurring and, in this case, based on his knowledge and experience, he estimates that there is a one-in-two chance of a boom and a 0.5 probability of a recession. Finally, the businessman estimates the financial consequences as an £80,000 profit for the new factory if there is a boom, and a £30,000 loss if there is a recession.

decision tree

an aid to decision-making in uncertain conditions, that sets out alternative courses of action and the financial consequences of each alternative, and assigns subjective probabilities to the likelihood of future events occurring. For, example, a firm thinking of opening a new factory the success of which will depend upon consumer spending (and thus the state of the economy) would have a decision tree like Fig. 32.

In order to make a decision, the manager needs a decision criterion to enable him to choose which he regards as the best of the alternatives and, since these choices involve an element of risk, we therefore need to know something about his attitudes to risk. If the manager were neutral in his attitude to risk then we could calculate the certainty equivalent of the ‘open factory’ alternative using the expected money value criterion, which takes the financial consequence of each outcome and weights it by the probability of its occurrence, thus:

which being greater than the £0 for certain of not opening the factory would justify going ahead with the factory project.

However, if the manager were averse to risk then he might not regard the expected money value criterion as being appropriate, for he might require a risk premium to induce him to take the risk. Application of a more cautious certainty equivalent criterion would reduce the certainty equivalent of the ‘open factory’ branch and might even tip the decision against going ahead on the grounds of the ‘downside risk’ of losing £30,000.See UNCERTAINTY AND RISK.

decision tree

a graphical representation of the decision-making process in relation to a particular economic decision. The decision tree illustrates the possibilities open to the decision-maker in choosing between alternative strategies. It is possible to specify the financial consequence of each ‘branch’ of the decision tree and to gauge the PROBABILITY of particular events occurring that might affect the consequences of the decisions made. See RISK AND UNCERTAINTY.
References in periodicals archive ?
With a high-risk population of 910 cases, a low-risk population of 299 cases, and an AUC of 0.629 (decision tree for detecting diabetes), the statistical power of the test was close to 100.0%.
A decision tree is built recursively following a top-down approach.
Decision Tree builds classification or regression models in the form of a tree structure.
Multilayer perceptron neural network and J48 decision tree machine learning algorithm and their individual performance will be recorded respectively and the output measured and evaluated.
Decision tree algorithms are applied for response and predictor variables in order to obtain homogenous subgroups depending upon sample size, nonlinear and the interaction effects of the predictor variables (Ali et al., 2015).
Examples of decision trees with fixed constraints are shown in Figure 9, while the associated solution vectors are shown in Table 5.
Characterization of variable importance measures derived from decision trees. Doctoral dissertation, Universite de Liege, Liege, Belgique.8-12.
Therefore, building decision trees with ID3 seemed to be a good starting point.
An examination of the decision tree between models 1 and 2 shows that nodes in model 2 are split two ways, indicating identical results to model 1.
As long as the decision trees in SW R offers possibility for reducing number of the variables by recursive feature elimination we use also wrapper feature selection algorithm, which in contrast to filter, exists as a wrapper around the induction algorithm.
The random forest using in this paper with probabilistic output is based on CART (Classification and Regression Tree) decision tree algorithm and the detailed steps of CART is showed at Table 1.
[23] China 344 : 344 70.9-92.1 Taiwan 220 : 220 Australian 307 : 382 German 700 : 300 MDA: multivariate discriminant analysis, DT: decision tree, NN: neural network, PCA: principal component analysis, LR: logistic regression, NB: Naive Bayes, MLP: multilayer perceptron neural network, CART: classification and regression tree, SVM: support vector machines, and RSBL: random subspace binary logit.

Full browser ?