Download 7 - WordPress.com

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Choice modelling wikipedia , lookup

Transcript
Model – Based Clustering Methods
Model Based clustering methods attempt to optimize the fit between the given data and
some mathematical model.
Model Based clustering methods follow two major approaches
1. Statistical approaches or conceptual clustering. &
2. Neural network approach.
Statistical Approach / Conceptual Clustering:
Statistical approach
in machine learning.
Given a set of unlabelled objects, it
produces a classification scheme over the objects. Apart from finding classes of like
objects, it also finds characteristic description of each class.
Most methods of conceptual clustering adopt a statistical approach that uses probability
measurements in determining the concepts or clusters. Probabilistic descriptions are
typically used to represent each derived concept.
COBWEB is a popular and simple method of incremental conceptual clustering. Its input
objects are described by categorical attribute-values pairs.
COBWEB uses a heuristic evolution measures called category utility to guide
construction of tree. Category utility(CU) is defined by
COBWEB descends the tree along an appropriate path, updating the counts (probability values)
along the way, the counts (probability values) along the way, in search of the best node at which
to classify the object. This decision is based on temporary placing the object in each node and
computing the “category utility” of the resulting partition. The placement that results in the
highest “category utility” is select as the node (class or concept) to incorporate the object.
COBWEB also computer the category utility of the partition / class that would result if a new
node were to be created for the object.If the object does not really belong to any of the concepts /
classes represented in the file so far, a new node is created for it and the object is incorporated
into it.
Neural Network Approach
Neural Network approach to clustering tends to represent each cluster as an “exemplar”.
An exemplar acts as a “prototype” of the cluster and does not necessarily have to correspond to a
particular data example or object.
Two prominent methods of the neural network approach
1. Competitive Learning
2. Self Organizing feature Maps (SOMs)
Competitive Learning:
It involves a hierarchical architecture of several units. That compete in a winner-takes-all for the
object that is currently being presented the system.