Tagged: GMM Toggle Comment Threads | Keyboard Shortcuts

  • wangxinxi 11:37 on January 30, 2014 Permalink | Reply
    Tags: GMM, ,   

    In general, you can avoid getting ill-conditioned covariance matrices by using one of the following precautions:

    Pre-process your data to remove correlated features.
    Set ‘SharedCov’ to true to use an equal covariance matrix for every component.
    Set ‘CovType’ to ‘diagonal’.
    Use ‘Regularize’ to add a very small positive number to the diagonal of every covariance matrix.
    Try another set of initial values.

    Advertisements
     
  • wangxinxi 21:43 on March 26, 2011 Permalink | Reply
    Tags: , GMM,   

    由于异常数据偏离mean值太远,导致GMM training的E-step过程中除0,进而出现nan。进而使得training过程失败。就是这个花了我两个小时去调试。

    问题是找出来了,怎么解决这个问题呢?

     
    • wangxinxi 01:58 on March 27, 2011 Permalink | Reply

      仔细分析后,才发现,原因在于EM算法的M-step中求导出了问题,以至于得到了一个不收敛的EM过程。虽然改进方法已经想到,但是为了简化问题,决定采用Univariate Gaussian.

  • wangxinxi 18:45 on March 21, 2011 Permalink | Reply
    Tags: , GMM   

    GMM can be used in graphical model to approximate a continuous density function. The derivation of the EM is so messy and time consuming -_-!. But finally, I got a quite nice closed form.

    The following is a useful link for matrix calculus in Gaussian distribution.
    http://www.cs.berkeley.edu/~pliang/cs281a/recitation-0924.pdf

     
c
Compose new post
j
Next post/Next comment
k
Previous post/Previous comment
r
Reply
e
Edit
o
Show/Hide comments
t
Go to top
l
Go to login
h
Show/Hide help
shift + esc
Cancel