Cluster-weighted modeling
In statistics, cluster-weighted modeling (CWM) is an algorithm-based approach to density estimation in joint input-output space proposed by Neil Gershenfeld.[citation needed] The basic CWM algorithm gives a single output cluster for each input cluster. However, CWM can be extended to multiple clusters which are still associated with the same input cluster.[1] Each cluster in CWM is localized to a Gaussian input region, and this contains its own trainable local model.[2] It is recognized as a versatile inference algorithm which provides simplicity, generality, and flexibility; even when a feedforward layered network might be preferred, it is sometimes used as a "second opinion" on the nature of the training problem.[3]
The original form proposed by Gershenfeld describes two innovations:
- Enabling CWM to work with continuous streams of data
- Addressing the problem of local minima encountered by the CWM parameter adjustment process[3]
CWM can be used to classify media in printer applications, using at least two parameters to generate an output that has a joint dependency on the input parameters.[4]
References
- ^ Feldkamp, L.A. (2001). "Cluster-weighted modeling with multiclusters" (PDF). International Joint Conference on Neural Networks. 3 (1): 1710โ1714.
{{cite journal}}
: Unknown parameter|coauthors=
ignored (|author=
suggested) (help) - ^ Boyden, Edward S. "1 Tree-based Cluster Weighted Modeling: Towards A Massively Parallel Real-Time Digital Stradivarius" (PDF). Cambridge, MA: MIT Media Lab.
{{cite journal}}
: Cite journal requires|journal=
(help) - ^ a b Prokhorov, A New Approach to Cluster-Weighted Modeling Danil V. "A New Approach to Cluster-Weighted Modeling". Dearborn, MI: Ford Research Laboratory.
{{cite journal}}
: Cite journal requires|journal=
(help); Unknown parameter|coauthors=
ignored (|author=
suggested) (help) - ^ Gao, Jun (2003-07-24). "CLUSTER-WEIGHTED MODELING FOR MEDIA CLASSIFICATION". Palo Alto, CA: World Intellectual Property Organization.
{{cite journal}}
: Cite journal requires|journal=
(help); Unknown parameter|coauthors=
ignored (|author=
suggested) (help)