|
|
Accuracy
| Name |
Accuracy |
| Description |
The Accuracy is a standard measure to evaluate the quality of hypotheses
in the context of concept learning.
Hypotheses are classifiers, determining
if instances of an underlying universe of discourse belong to a specific concept or
not.
Algorithms addressing the task of concept learning usually output such a hypothesis
after processing a given training set. A training set consists of classified instances,
i.e. it is given if instances belong to the target concept or not.
To estimate the quality of such a hypothesis, another set, a so called test set,
is separated from the set of available (classified) examples, and the learning algorithm
is only provided with the remaining examples. Afterwards the (yet unseen) examples of the
test set are classified by the hypothesis and the predicted class of each instance is compared
to the correct one. The accuracy is the ratio of correct classified instances to number of instances
in the test set.
|
| Translations |
Korrektheit (German)
|
|
|
