|
Home
|
Sign up!
|
Projects
|
Seminars
|
Research Notes
|
Today's Lecture
|
About
|
Update Research Note Form
Course:
Research Note Topic:
Research Note Description:
Ever thought how google and other search engines give you the best result everytime? Of course, they classify the millions of webpages according to some criteria and produce the results according to the user's request. But, ever thought how tedious it could be to do that?
In order to classify data in a more dynamic manner, computer systems try to classify data (either webpages or Genetic data) into various classes. Many algorithms are used for this type of classification -
Clustering Analysis
,
Perceptrons
(Neural Networks),
MaxSim's Formula
,
SVM
(Support Vector Machine),
SBC
(Similarity Based Classification),
HMM
(Hidden Markov Model)
etc.....
Though almost all of these algorithms are used very successfully to classify data, much is still desired from classification algorithms. For example:
Every classification algorithm treats each set of data differently, i.e. the classifications by two algorithms of the same data may not be similar,
There is no set standards on what kind of algorithm will work on what kind of data. It is only a matter of hit and try to figure out the "best fit", and
All algorithms need a "decision parameter" with which they decide on the boundry seperating the various classes. This parameter has to be user defined and takes a lot of
tweaking
.
Your Password:
Prof. Ashay Dharwadker