how to average the weights

Oct 21, 2009 at 8:50 AM

Vertice1  Vertice2                 weights(frequency of communication)

A           B                             3

A           C                             2

B           A                              1

 

Above is the information I have for analysis in NodeXL. Here relationship is undirected..when I put this data on Nodexl and use merge function I found that weights for AB become 4 (not 2). As this is undirected relationship, weights between A and B should 2. Do you know any function of Nodexl to calculate in this way?

 

MAny thanks in advance..

 

Oct 21, 2009 at 5:06 PM

If you're merging an edge that has weight 3 with a duplicate edge that has weight 1, why would the resulting weight be 2?

-- Tony

Oct 21, 2009 at 8:51 PM

I just noticed the "average" in your discussion title.  NodeXL's merge feature adds edge weights, it doesn't average them.  And I can't think of a way to tell Excel to group and average dissimilar rows.  Perhaps someone else knows a good technique.

-- Tony

Nov 6, 2009 at 8:43 AM

Tony thanks a lot for your reply and sorry for my late response..!! I am a beginner in using this program. But quite fascinated by this tool and this kind of analysis :)

 

 

If you're merging an edge that has weight 3 with a duplicate edge that has weight 1, why would the resulting weight be 2?

--> As in my analysis(based on survey outcomes), we asked A B respectively on the perceived frequency of communication (A-B) for the same period. 2, 4 means they remember differently. Thus, I need to average them out as the closest proxy for A-B communication frequency.

 

I just noticed the "average" in your discussion title.  NodeXL's merge feature adds edge weights, it doesn't average them.  And I can't think of a way to tell Excel to group and average dissimilar rows.  Perhaps someone else knows a good technique.

Okay, I think I will just do this manually. But thanks a lot for sharing your idea..