Two commonly used data smoothing algorithms __ algorithm

Two kinds of data smoothing algorithms commonly used

Description: Laplace smoothing algorithm is a reference to the Chinese Academy of Sciences, Wang Bin's Modern information retrieval ppt-Turing (good-turing) smoothing algorithm turn from: http://hi.baidu.com/kangwp/item/8533124292d026e6bdf45 150

The smoothing algorithm is maxi.

First, Laplace smooth

Also known as Add 1 smoothing, now for text processing as an example:

, which is the selected number of features.

Second, the ancient German-Turing (good-turing) Smooth

The Good-Turing (good-turing) estimation method is at the heart of many smoothing techniques, and was proposed in 1953 by I.j.good, citing Turing (Turing). The basic idea is to use frequency category information to smooth the frequency. For any n-ary syntax where r times occur, it is assumed that it has occurred r* times.