VMEAN is the jVarfun module of jLab.

 VMEAN  Mean over non-NaN elements along a specified dimension.
 
    Y=VMEAN(X,DIM) takes the mean of all non-NaN elements of X along      
    dimension DIM. 
                                                                          
    [Y,NUM]=VMEAN(X,DIM) also outputs the number of non-NaN data points 
    NUM, which has the same dimension as X.              
 
    [Y1,Y2,...YN]=VMEAN(X1,X2,...XN,DIM) or also works, where all the XN
    are the same size.  
 
    VMEAN(X1,X2,...XN,DIM);  with no output arguments overwrites the 
    original input variables.
    __________________________________________________________________
 
    Weighted means
 
    VMEAN can also form a weighted mean.
 
    Y=VMEAN(X,DIM,W) forms the weighted mean of X using weights W, an array 
    of the same size as X. 
 
    In this case [Y,NUM]=VMEAN(X,DIM,W) returns total weight used in 
    forming the mean at each point, rather than the number of data points.
   
    Thus VMEAN(X,DIM,W) is the same as VMEAN(X.*W,DIM)./VMEAN(W,DIM).
 
    [Y1,Y2,...YN]=VMEAN(X1,X2,...XN,DIM,W) also works. 
    __________________________________________________________________
    This is part of JLAB --- type 'help jlab' for more information
    (C) 2001--2015 J.M. Lilly --- type 'help jlab_license' for details