Model-Based ¾ÛÀà¼ò½é
¾ÛÀàÎÊÌâµÄÁíÒ»Àà·½·¨ÊÇ»ùÓÚÄ£Ð͵ķ½·¨£¬¼´Ê¹ÓÃÌØ¶¨µÄÄ£Ð;ÛÀ࣬²¢ÊÔͼÓÅ»¯Êµ¼ÊÊý¾ÝÓëÄ£Ð͵ÄÊÊÅä¶È¡£
ʵ¼ÊÉÏ£¬Ã¿¸ö·ÖÀà¶¼¿É±í´ïΪÊýѧÉϵIJÎÊý·Ö²¼£¬Èç¸ß˹·Ö²¼£¨Á¬Ðø£©»ò²´ËÉ·Ö²¼£¨ÀëÉ¢£©¡£Òò´ËÕû¸öÊý¾Ý¼¯Ôò¿É½¨Ä£³ÉÒ»¸öÕâЩ·Ö²¼µÄ»ìºÏmixture·Ö²¼£¬½¨Ä£Ìض¨·ÖÀàµÄµ¥¸ö·Ö²¼¾³£±»³ÆÎª component·Ö²¼¡£
Ò»¸ö»ìºÏÄ£Ðͼ«ÓпÉÄÜÓÐÏÂÁÐÌØµã£º
- component ·Ö²¼ÓС±¸ß·å¡±£¬ (µ¥¸ö·ÖÀàµÄÊý¾Ý½ôÃܽӽü£©;
- mixtureÄ£Ðͺܺõء°¸²¸Ç¡±Êý¾Ý (ÓÉÓÚcomponent·Ö²¼ºÜºÃµØ±í´ïÁËÊý¾Ý·Ö²¼Ä£Ê½).
»ùÓÚÄ£Ð͵ľÛÀàÖ÷ÒªÓŵãÈçÏ£º
- ¿ÉÒÔʹÓóÉÊìµÄͳ¼ÆÍƶϼ¼Êõ;
- Áé»îÑ¡Ôñcomponent distribution;
- ¿ÉÒԵõ½Ã¿¸ö·ÖÀàµÄ¹À¼ÆÃܶÈ;
- ¿ÉÒÔÓÃʹÓÃÒ»¸ö ¡°Èí¡± ·ÖÀà
¸ß˹»ìºÏ
»ù±¾Ë¼Ï룺¼ÙÉèÕû¸öÊý¾Ý¼¯·þ´Ó¸ß˹»ìºÏ·Ö²¼£¬´ý¾ÛÀàµÄÊý¾Ýµã¿´³ÉÊÇ·Ö²¼µÄ²ÉÑùµã£¬Í¨¹ý²ÉÑùµãÀûÓÃÀàËÆ¼«´óËÆÈ»¹À¼ÆµÄ·½·¨¹À¼Æ¸ß˹·Ö²¼µÄ²ÎÊý¡£Çó³ö²ÎÊý¼´µÃ³öÁËÊý¾Ýµã¶Ô·ÖÀàµÄÁ¥Êôº¯Êý¡£
ÕâÀà·½·¨ÖУ¬Ê¹ÓÃ×î¹ã·ºµÄÊǸß˹»ìºÏÄ£ÐÍ£ºÎÒÃÇʵ¼ÊÉÏ¿ÉÒÔ°Ñ·ÖÀà¿´³ÉÒÔÖØÐÄΪÖÐÐĵĸß˹·Ö²¼£¬ÈçÏÂͼ£¬»ÒɫԲȦ±íʾ·Ö²¼µÄÖ÷Òª²¿·Ö£º

Ëã·¨¹¤×÷ÈçÏ£º
- ÒÔ¸ÅÂÊ
Ëæ»úÑ¡ÔñÒ»¸öcomponent (the Gaussian) ·Ö²¼ ; - ´Ó·Ö²¼
ÖвÉÑùÒ»¸öµã.
Let¡¯s suppose to have:
- x1, x2,..., xN

We can obtain the likelihood of the sample:
.
What we really want to maximise is
(probability of a datum given the centres of the Gaussians).

is the base to write the likelihood function:

Now we should maximise the likelihood function by calculating
, but it would be too difficult. That¡¯s why we use a simplified algorithm called EM (Expectation-Maximization).
The EM Algorithm
The algorithm which is used in practice to find the mixture of Gaussians that can model the data set is called EM (Expectation-Maximization) (Dempster, Laird and Rubin, 1977). Let¡¯s see how it works with an example.
Suppose xk are the marks got by the students of a class, with these probabilities:
x1 = 30 
x2 = 18 
x3 = 0 
x4 = 23 
First case: we observe that the marks are so distributed among students:
x1 : a students
x2 : b students
x3 : c students
x4 : d students

We should maximise this function by calculating
. Let¡¯s instead calculate the logarithm of the function and maximise it:

Supposing a = 14, b = 6, c = 9 and d = 10 we can calculate that
.
Second case: we observe that marks are so distributed among students:
x1 + x2 : h students
x3 : c students
x4 : d students
We have so obtained a circularity which is divided into two steps:
- expectation:

- maximization:

This circularity can be solved in an iterative way.
Let¡¯s now see how the EM algorithm works for a mixture of Gaussians (parameters estimated at the pth iteration are marked by a superscript (p):
- Initialize parameters:
 - E-step:
- M-step:
where R is the number of records.
|
ÉÏÎĺóÃæÃ»Ì«¿´¶®£¬ÕâÀï´ÓcsdnµÄÒ»¸ö²©¿Íժ¼ÁËһЩ¸ß˹»ìºÏ¾ÛÀàËã·¨£¬½²µÄ»¹±È½ÏÇå³þ£¬
¼ÙÉèn¸ö¸ß˹·Ö²¼£¬´ÓÖвúÉúm¸öÑù±¾
,ÿһ¸ö yj Ëù´ÓÊôµÄ·Ö²¼Óà zj ±íʾ£¬zj µÄȡֵ·¶Î§Îª1µ½n¡£¶ÔÓÚÈÎÒâµÄy£¬ÆäÀ´×ÔµÚi¸ö¸ß˹·Ö²¼µÄ¸ÅÂÊΪ

ÎÒÃǵÄÈÎÎñÊ**À¼ÆÎ´ÖªµÄ²ÎÊý
£¬¼´Ã¿Ò»¸ö¸ß˹·Ö²¼µÄ¾ùÖµ¡¢·½²îºÍÏÈÑé¸ÅÂÊ¡£
Ê×ÏÈÎÒÃDZØÐëÉ趨³õʼֵ¡£
¶ÔÓÚÎ޼ලµÄ¾ÛÀ࣬ÎÒÃÇÊ×ÏȱØÐëΪÿһ¸öÑù±¾Ö¸¶¨Ò»¸öÀà±ð£¬ÕâÒ»²½¿ÉÒÔͨ¹ýÆäËûµÄ¾ÛÀà·½·¨ÊµÏÖ£¬ÈçK-means·½·¨£¬Çó³ö¸÷¸öÀà±ðµÄÖÐÐĺÍÿһ¸öÑù±¾µÄÀà±ð¡£È»ºóÇó³ö¸÷¸öÀà±ðÖÐÑù±¾µÄз½²îÕ󣬿ÉÒÔÓÃÿ¸öÀà±ðÖÐÑù±¾µÄ¸öÊýÀ´±íʾ¸ÃÀà±ðµÄÈ¨ÖØ£¨ÏÈÑé¸ÅÂÊ£©¡£
E-step
ʹÓÃÉÏÒ»´ÎM²½Öеõ½µÄ²ÎÊý¦È£¬¼ÆËãδ֪µÄz£¨¼´Ã¿Ò»¸ö·ÖÀࣩ¶ÔÓÚ¹Û²âÊý¾ÝµÄÌõ¼þ·Ö²¼£¬¼´£¬ÔÚµ±Ç°²ÎÊý¦ÈnÏ£¬¶Ôÿһ¸öÑù±¾£¬¼ÆËã¸÷¸öÀà±ðµÄºóÑé¸ÅÂÊ¡£ ÏÂʽÖеÚÈý¸öµÈºÅºóÃæµÄ·Ö×ӵĵÚÒ»ÏΪÿһ¸öÑù±¾ÔÚÀà±ðiÖеķֲ¼¸ÅÂÊ£¬·Ö×ÓÖеĵڶþÏîΪÿһ¸öÀà±ðµÄÏÈÑé¸ÅÂÊ¡£
·ÖĸΪ¹éÒ»»¯£¬¼´¶Ôÿһ¸öÑù±¾£¬ÆäÔÚ¸÷¸öÀà±ðÖеķֲ¼¸ÅÂʵĺÍΪ1.

for ti = 1:Ncluster
PyatCi = mvnpdf(featurespace,Mu(ti,:),Sigma(:,:,ti));
E(:,ti) = PyatCi*Weight(ti);
end
Esum = sum(E,2);
for ti = 1:Ncluster
E(:,ti) = E(:,ti)./Esum;
end
M-step
![¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà ¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà]()
![¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà ¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà]()
![¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà ¾ÛÀàËã·¨½Ì³Ì£¨4£©£º¸ß˹»ìºÏ¾ÛÀà]()
ΪÁË×î´ó»¯ÁªºÏ·Ö²¼µÄ¶ÔÊýËÆÈ»º¯ÊýµÄÆÚÍû
ÉÏʽÖÐ×îºóÒ»ÏîµÄÁªºÏ·Ö²¼¿ÉÒÔд³ÉÌõ¼þ·Ö²¼ºÍ±ßÔµ·Ö²¼µÄ³Ë»ýµÄÐÎʽ£¬µÃµ½

ÉÏʽÐèÒªÂú×ãÏÈÑé¸ÅÂʹéÒ»»¯µÄÒªÇó£º

ΪÁËÇóµÃQ(¦È)µÄ¼«´óÖµµÄλÖã¬ÒýÈëÀ¸ñÀÊÈÕËã×Ó £¬È»ºó´øÈë¸ÅÂÊÃܶȺ¯Êý£¬

ΪÁËÇóµÃÐ嵀 ¦Èt + 1,ÎÒÃÇÐèÒªÕÒ³öÉÏʽȡµÃ¼«ÖµÊ±¶ÔÓ¦µÄ¦È£¬¼´ÆäÒ»½×µ¼Êý
¡£
ÏÂÃæÍÆµ¼Âú×ãÉÏʽµÄ¾ùÖµ£¬

for ti = 1:Ncluster
muti = E(:,ti)'*featurespace;
muti = muti/sum(E(:,ti));
Mu(ti,:) = muti;
end
²úÉúеÄз½²î¾ØÕó
еĸ÷¸ö·ÖÀàµÄÈ¨ÖØ
½øÐйéÒ»»¯´¦Àí
Inserting ¦Ë into our estimate:
½«ÐµIJÎÊý£¬¾ùÖµ¡¢·½²î¡¢·ÖÀàµÄÏÈÑé¸ÅÂÊ£¬´øÈëE²½£¬ÖØÐ¼ÆË㣬ֱµ½Ñù±¾¼¯ºÏ¶Ô¸÷¸ö·ÖÀàµÄËÆÈ»º¯Êý²»ÔÙÓÐÃ÷ÏԵı仯Ϊֹ¡£
Bibliography