Date: Thu, 18 Dec 1997 11:52:29 -0500 (EST)
From: "Allison L. Powell"
To: Bei Wu, Travis Emmitt
Cc: "James C. French"
Subject: Jaccard matrix

All of the data for this project lives on an attached disk on
the compute server altair.  So, you'll need to log onto altair to do
this work.  The file that you're working on is
/usr/ads/dice/matrix.jaccard

The format is as follows:

row_id1
col_id1 jaccard1 col_id2 jaccard2 col_id3 jaccard3

For example:

D1024
D1023 0.4000 D1022 0.3636 D1018 0.1667 D1016 0.4000 D1014 0.2727 D1013 0.2143 D1012 0.2143 D1009 0.2857 D1007 0.2308
D1023
D1022 0.6250 D1018 0.5000 D1016 0.5000 D1014 0.5000 D1013 0.3636 D1012 0.3636 D1009 0.4545 D1007 0.4000
D1022
D1018 0.3000 D1016 0.3000 D1014 0.4444 D1013 0.3333 D1012 0.3333 D1009 0.3077 D1007 0.3636
D1021
D1020 0.6250 D1002 0.1818 D1001 0.1818
D1020
D1002 0.2000 D1001 0.2000
D1018
D1016 0.5000 D1014 0.3333 D1013 0.2500 D1012 0.2500 D1009 0.3333 D1007 0.2727

The above example represents six rows of the matrix.

The values that start with a 'D' are names for different affiliation
strings and are used to name rows and columns in this representation
of the matrix.  Do NOT make any changes to these names, only change the
jaccard values to dice values.

At the end of the file are several entries of the form:

D4931

D1242

D2317

D1196

D5016

This means that no other strings were sufficienly similar to these
(i.e. all zero entries).  Do not change any of these items.

Just let me know if you have any questions or problems.

Allison