Changes between Version 34 and Version 35 of adherelinux/cocluster


Ignore:
Timestamp:
Sep 8, 2010, 2:01:05 PM (14 years ago)
Author:
adherelinux
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • adherelinux/cocluster

    v34 v35  
    5454定義[[Image(W.jpg)]],上標的T為轉置符號[[br]],然而W矩陣如下[[Image(wij.jpg)]](8),D表示為document-word weight matrix
    5555[[Image(D.jpg)]](9),注意W和D有相同的列跟行,他可以表示出degree of aggregation(聚合度)公式如下[[br]]
    56 [[Image(sumuvd.jpg)]],當(7)是當成物件函數時,其中在W中的各個[[Image(w.jpg)]]元素都會小於或等於1,在這裡我們考慮兩個membership vectors[[br]][[Image(uT.jpg)]] and [[Image(VT.jpg)]].基於(8)式子我們得知[[Image(sumuv.jpg)]],其中[[Image(|UT|.jpg)]] [[br]]
     56[[Image(sumuvd.jpg)]](10),當(7)是當成物件函數時,其中在W中的各個[[Image(w.jpg)]]元素都會小於或等於1,在這裡我們考慮兩個membership vectors[[br]][[Image(uT.jpg)]] and [[Image(VT.jpg)]].基於(8)式子我們得知[[Image(sumuv.jpg)]],其中[[Image(|UT|.jpg)]] [[br]]
    5757由於Ruspini`s condition  [[Image(sumu=1.jpg)]]所以可以得知[[Image(U<=1.jpg)]],同樣的[[Image(v<=1.jpg)]]and[[Image(cos.jpg)]].因此[[Image(wij<=1.jpg)]]
     58這邊我們考慮W為單位矩陣,例如全部的元素都為1,由式子(10)我們可以計算出 degree of aggregation為
    5859
    59