Pattern Recognition and Application by Prof. P.K. Biswas,Department of Electronics & Communication Engineering,IIT Kharagpur.For more details on NPTEL visit nptel.ac.in
The question is old but someone may still be interested: It's not: The hyperplane dividing the classes has the same dimension of the feature space. If you see he is using a line (dimension = 2 ) to divide examples on a 2d space. If you want to classify a set of gym athletes as Tall or Short based only on their height you can plot the height values on a 1D space and use a point (1D boundary) that acts like a threshold and divides the classes Tall and Short. You just have to propagate this reasoning to a higher dimensional space
hyperplane will have the same dimension as vector space but the difference is that, if in vector space you have n independent variable then in hyperplane you will have n-1 independent variable.
Bayes Decision Theory 32:00
Probability of error 50:44
Amazing lecture. Simply explained this difficult topic. Salute to you Sir.
Watch from 32:00 for Bayesian Decision Theory.
Best video best teacher. lots of respect !
Graet Sir
Very helpful ,Thanks for uploading.
Very helpful. Thank you very much.
Thank you very much professor.
very helpful sir.
very well-explained!
Isn't the dimension of the hyper plane one less than the dimension of the vector space defining the pattern ?
I agree with you
same doubt
The question is old but someone may still be interested:
It's not:
The hyperplane dividing the classes has the same dimension of the feature space.
If you see he is using a line (dimension = 2 ) to divide examples on a 2d space.
If you want to classify a set of gym athletes as Tall or Short based only on their height you can plot the height values on a 1D space and use a point (1D boundary) that acts like a threshold and divides the classes Tall and Short.
You just have to propagate this reasoning to a higher dimensional space
hyperplane will have the same dimension as vector space but the difference is that, if in vector space you have n independent variable then in hyperplane you will have n-1 independent variable.
At 51:12, shouldn't the vertical axis be labeled P(x|w) ? Otherwise, the points along the graphs at any given x would have to sum to 1.
can we say evidence that is (p(x)) can be considered as scalable factor for determining the class?
Nice lecture on this topic but sound is very poor
Bayes decision theory 31.00
watch in 1.25
Very low sound
bhot zyada volume h kam kro thoda