comp.lang.idl-pvwave archive: archive » how to calculate mcnemar statistics based on confusion matrix

Home » Public Forums » archive » how to calculate mcnemar statistics based on confusion matrix

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

how to calculate mcnemar statistics based on confusion matrix [message #84940]

Wed, 19 June 2013 05:26

pulchowk-campus-alumn is currently offline

pulchowk-campus-alumn
Messages: 2
Registered: January 2012

Junior Member

Hi all,
I am trying to compare accuracies using two different algorithms using same training data. I thought McNemar test would be good option for this (http://en.wikipedia.org/wiki/McNemar's_test).

Those people who are familiar with that could you please elaborate me how to calculate it based on the confusion matrices that i got using two different algorithms.

My samples for different classes:
23711
6212
12219
8149
2874
5787
3169
6569

Confusion matrix from algo 1:
20867 401 8 966 0 131 171 1167
8 5400 116 263 2 423 0 0
0 0 12213 0 0 0 6 0
1734 80 30 5532 0 319 15 439
0 36 48 0 2681 109 0 0
36 99 80 444 0 5051 67 10
0 0 30 0 0 11 3104 24
1054 103 82 1540 2 809 731 2248

Confusion matrix from algo 2:
21166 40 14 435 0 226 231 1599
166 5637 134 103 0 155 0 17
55 1 12151 0 0 0 12 0
3019 49 38 4287 0 203 34 519
5 76 2 26 2508 252 5 0
387 131 66 211 0 4865 95 32
4 0 10 3 0 8 3135 9
1346 96 68 1849 0 354 776 2080

Now, how to calculate mcnemar statistics out of it.

Report message to a moderator

Previous Topic:	convert byte array [48,1221] to float
Next Topic:	Interpolation

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 13:33:27 PDT 2025

Total time taken to generate the page: 0.00396 seconds