Article Directory
"Naive" Origin
Naive Bayes (Naive Bayes) is a group of supervised learning algorithm, which is based on Bayes' theorem, "naively" "naive" is assumed independent of each other between the features, i.e., the presence of a characteristic independent of the presence of other features. This is the "simple" reason the word "Naive".
Bayes's law and case: https://blog.csdn.net/houhuipeng/article/details/90706539
sample
Belong to the category
probability:
When satisfied
is
's
Probability greater than Example
When the embodiment, may be determined
belong
is assumed mutually independent features, then
finally available:
Advantages and disadvantages
Naive Bayes model is a set of extremely quick and simple classification algorithm, usually for very high-dimensional data sets. Because they are so fast and tunable small, so they end up as a very useful basis for rapid classification problems.
Advantages: speed operation, less simple model tunable parameters.
Cons: assumption is not true in most cases.
Depending on the assumptions feature distribution can be divided into:
- Naive Bayesian Gaussian
- Naive Bayes polynomial
- Supplementary Naive Bayes
- Bernoulli Naive Bayes
Participate sklearn official document:
https://scikit-learn.org/stable/modules/naive_bayes.html