-
Even the feature are dependent.
- relativities score(prob)
- positive correlated --> over confident --> relativities score(prob)
- negtive correlated --> less contributiion for
P(C|F)
-
assume independence
- easier to implement / parallelize --> can compute large of data amounts
-
when to use
- high cardinality (url, user, ...)
- eager learning (realtime learning) by maintain freq tables.
releted paper
[1] C. D. Manning, P. Raghavan, H. Schutze, An introduction to information retrieval, Cambridge University Press, April 2009.
[2] H. Zhang, The optimality of naive Bayes, American Association for Artificial Intelligence, 2004.
[3] S. S. Y. Ng, Y. Xing and K. L. Tsui, A naive Bayes model for robust remaining useful life prediction of lithium-ion battery, Applied Energy, 118, p. 115, 2014.
Post navigation