classification - Does MLE produce a generative or discriminative classifier? -


i understand general concept of generative versus discriminative classifiers: generative classifier can produce actual model of data. discriminative classifiers on other hand produce decision.

the problem have concept generative classifier needs prior in order produce (generate) data. if parameter produced mle sufficient reproduce data, isn't a generative model, @ least criteria above.

the problem having, think, if parameters of model not known, distribution of parameters required model meaningful. however, given data sufficient produce mle of paramters, @ point parameters taken given , becomes generative model.


Comments

Popular posts from this blog

jsf - "PropertyNotWritableException: Illegal Syntax for Set Operation" error when setting value in bean -

java - How to read a file created at runtime? -

php - Autoloader issue not returning Class -