Predictive Classification Function
(Redirected from predictive classification function)
Jump to navigation
Jump to search
A predictive classification function is a classification function that is a predictive function.
- Context:
- input: a Test Instance Set (of test instances).
- range: It produces one or more class predictions (which can be Correct Prediction or Incorrect Prediction).
- Performance Measures: such as Accuracy.
- It can range from being a Binary Classification Model to being a Multiclass Classification Model.
- It can range from being a Discriminative Classification Model to being a Generative Classification Model.
- It can be produced by a Classifier Learning System.
- It can range from being an Overfitted Predictive Classifier to being a (well) Generalized Predictive Classifier.
- Its Performance can be visualized with a Confusion Matrix (presents how well the classifier predicts each of the available classes)
- If a Cost-Benefit Matrix is available then the classifier's performance is measured by the product of the Confusion Matrix and Cost-Benefit Matrix.
- It can be an input/output in a Classification Task.
- Example:
- Decision Tree Classifier, Neural Network Classifier, Nearest Neighbor Classifier, Support Vector Machine Classifier, Bayes Network Classifier, Markov Random Field Classifier.
- Domain-Specific Predictive Classification Function, such as:
- A GrantLoan(Person) =>Y/N classification model predicts whether a person belongs in the Concept Class of good credit risk people.
- A Named Entity Classifier predicts whether a noun phrase is of a certaion Named Entity type. E.g. PERSON, LOCATION, DATE, etc.
- A Relation Recognition Classifier predicts what types of Semantic Relations exists in a Sentence.
- …
- Counter-Example(s):
- See: Classification Model Evaluation Metric, Generalization.
References
2009
- (WordNet, 2009) ⇒ http://wordnetweb.princeton.edu/perl/webwn?s=classifier
- S: (n) classifier (a person who creates classifications)
- S: (n) classifier (a word or morpheme used in some languages in certain contexts (such as counting) to indicate the semantic class to which the counted item belongs)
- http://www.predictionworks.com/glossary/
- Classifier: A model that classifies is sometimes referred to as a "classifier". Commonly a classifer's performance is measured by its ability to correctly label unseen test cases, that is its "accuracy". Inversely a classifier's performance may be measured by its "error rate". A more detailed insight into a classifier's performance is given by the Confusion Matrix structure because it captures how well the classifier predicts each of the available classes. If a Cost-Benefit Matrix is available then the classifier's performance is measured by the product of the Confusion and Cost-Benefit matrices. See also: Accuracy, Classification Algorithm, Confusion Matrix, Cost-Benefit Matrix, Estimation, Model, and Type I and Type II Errors.
2008
- Peter L. Rosenbaum, Robert J. Palisano, Doreen J. Bartlett, Barbara E. Galuppi, and Dianne J. Russell. “Development of the gross motor function classification system for cerebral palsy." Developmental Medicine & Child Neurology 50, no. 4 (2008): 249-253.
1998
- (Kohavi & Provost, 1998) ⇒ Ron Kohavi, and Foster Provost. (1998). “Glossary of Terms.” In: Machine Leanring 30(2-3).
- Classifier: A mapping from unlabeled instances to (discrete) classes. Classifiers have a form (e.g., decision tree) plus an interpretation procedure (including how to handle unknowns, etc.). Some classifiers also provide probability estimates (scores), which can be thresholded to yield a discrete class decision thereby taking into account a utility function.
1995
- (Kohavi, 1995) ⇒ Ron Kohavi. (1995). “A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection.” In: Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence (IJCAI 1995).
- QUOTE: A classifier is a function that maps an unlabelled instance to a label using internal data structures. An inducer or an induction algorithm builds a classifier from a given dataset. CART and C4.5 (Brennan, Friedman Olshen & Stone 1984, Quinlan 1993) are decision tree inducers that build decision tree classifiers. In this paper we are not interested in the specific method for inducing classifiers, but assume access to a dataset and an inducer of interest.