```Date: Wed, 1 Nov 2006 11:29:45 -0800 Reply-To: Ruve Sender: "SAS(r) Discussion" From: Ruve Organization: http://groups.google.com Subject: How to infer the UNKNOWN cells Comments: To: sas-l@uga.edu Content-Type: text/plain; charset="iso-8859-1" Hi, All Here is a statistical problem, hopefully. I have a table that shows customers' financial performance at the segment level. All customers are segmented by 2 dimensions - group and cat. The cat dimension is ordinal and the group dimension is nominal (but can be converted to ordinal). The values reflect the metric under study. Now the problem is that we do not have any observations for cat1-cat4 or group21-group25. Is there a way to infer the missing cells based on the known values? Thanks! cat1 cat2 cat3 cat4 cat5 cat6 cat7 cat8 cat9 group1 4.57 9.27 15.65 39.22 45.45 group2 4.99 13.93 20.83 47.73 36.66 group3 8.26 14.51 38.07 71.40 128.00 group4 6.21 11.15 27.23 74.49 93.42 group5 4.57 9.29 23.43 73.09 91.78 group6 6.61 14.38 32.59 86.40 125.57 group7 10.57 9.78 16.96 33.00 21.18 group8 7.95 7.87 11.41 27.74 86.56 group9 9.22 5.84 14.20 27.92 22.19 group10 6.27 49.21 19.46 49.39 97.72 group11 6.86 7.73 15.51 53.01 118.98 group12 33.57 16.17 24.63 73.64 124.77 group13 16.47 11.19 23.94 91.16 92.08 group14 6.87 10.58 20.95 169.51 59.66 group15 4.51 14.07 19.14 67.71 36.81 group16 9.72 16.40 25.38 71.45 139.66 group17 6.19 11.75 13.89 9.76 67.41 group18 4.67 11.24 12.87 3.50 6.95 group19 1.94 3.50 5.32 5.55 25.64 group20 3.67 4.11 8.60 15.38 32.72 group21 group22 group23 group24 group25 ```

Back to: Top of message | Previous page | Main SAS-L page