User:Laura martignon/sandbox

= Fast and frugal trees =

Green and Mehr (1997) constructed what is now called a fast and frugal tree for decisions as to whether a patienz with heart disease has to be assigned to the Coronary Care Unit (CCU). The fast and frugal tree works like this: if a patient has a certain anomaly in his electrocardiogram — the ST segment (a particular segment of the ECG) is elevated — he is immediately classified as being at a high risk and will be assigned to the CCU. No other information is searched for. If that is not the case, a second question is asked: Is the patient’s primary complaint chest pain? If this is not the case, he is classified as low risk and assigned to a regular nursing bed. No further information is considered. If the answer is yes, a final question, namely whether any of four remaining symptoms beyond chest pain is present or not, is asked to classify the patient and decide on the appropriate treatment, namely use of nitroglycerine for relief (NTG), patient'S history of microcardial infarction (MI), ST segment flattening and hyperacute or inverted T-waves. The tretion below represents the classification process.

{Baum als categorization tree: low risk/high risk)

Description
Fast and frugal trees are a specific class of heuristics for binary categorization. They can be described in terms of three building blocks: (1) ordered search, (2) fast stopping rule, and (3) one-reason decision-making. More specifically, a fast and frugal tree is a binary categorization tree with at least one exit leaf at each level and two exit leaves at the very last level. A level corresponds to a “cue for categorization” in a risky world. No interdependencies between features are taken into account, and the features are ranked according to a specific criterion of “goodness for categorization”. For instance the positive and negative validity can be used, or sensitivity and specificity [links!]. Fast and frugal trees for categorization become fast and frugal decision trees when each category is associated with a binary decision. For instance such a tree can be applied for deciding whether to prescribe antibiotic treatment to young children suffering from community acquired pneumonia. The following tree describes this decision process.

Theory
Fast and frugal trees are a special case of Simple Heuristics [link!] — simple decision strategies that are surprisingly robust in prediction, even when compared with more complex methods (Gigerenzer, Todd, & the ABC research group, 1999; Gigerenzer, Hertwig, & Pachur, 2011) — and have been linked with the theoretical framework for diagnostic classification decisions provided by signal detection theory (SDT, Luan, Schooler, & Gigerenzer, 2011). Specifically, the exit structure of a fast and frugal tree can be rearranged to achieve a certain kind of decision bias. For example, the tree in Green and Mehr tends to send more people to CCU (coronary care unit) than regular nursing bed relative to their base rates. If the first exit in the tree is changed from pointing to the right to pointing to the left, then even more people will be sent to CCU. On the contrary, if the second exit in the tree is changed to a different direction (while other exits stay the same), then more people will be sent to Regular Nursing Bed. This characteristic enables a person to apply fast and frugal trees adaptively in tasks with different payoff structures that require different kinds of decision biases in order to get better payoffs. In other words, it makes fast and frugal trees more ecologically rational (link to the wiki entry).

[hier schreibt Laura über Mathematik]

Applications
Here comes more text