Yes, we got your first email.
We have not had a chance to look at the issue.
The underlying (fortran) code has had some problems.
My suggestion at this time is to see if you can find another
tool or code to perform the canonical correlation.
Sorry.
On 4/13/10 9:57 PM, sidong7041@hotmail.com wrote:
> Hi there,
> I am performing a canonical correlation analysis with the cancor routine.
> The two fields is SST( 158,55) and rain( 156,55).
> however, I got the following incorrect results. The output canr value is
> incorrect （large than 1.0）.
> Thanks a lot in advance,
> --Stone
> --------------------------------------------------------------------------
> The output :
> canr
> Dimensions and sizes: [156]
> Coordinates:
> Number Of Attributes: 5
> ndof : <ARRAY of 156 elements>
> coefy : <ARRAY of 24336 elements>
> coefx : <ARRAY of 24648 elements>
> wlam : <ARRAY of 156 elements>
> chisq : <ARRAY of 156 elements>
> (0) 177156.9
> (1) 143534.9
> (2) 126688.5
> (3) 118556.6
> (4) 111030.6
> (5) 106364
> (6) 97091.65
> (7) 96355.91
> (8) 90402.44
> (9) 85798.05
> (10) 81377.4
> (11) 80841.88
> (12) 77203.06
> (13) 73328.29
> (14) 71752.25
> (15) 67675.48
> (16) 65583.56
> (17) 65149.32
> (18) 61743.71
> (19) 59880.7
> (20) 56873.12
> (21) 55540.66
> (27) 46672.02
> (28) 45966.54
> (29) 45708.81
> (30) 44472.45
> (31) 42724.57
> (32) 42007.14
> (33) 41123.85
> (34) 39644.17
> (35) 38480.35
> (36) 36747.88
> (37) 36268.79
> (38) 34981.12
> (39) 33899.27
> (40) 32424.7
> (41) 31713.69
> (42) 30990.04
> (43) 30100
> (44) 28837.65
> (45) 28029.51
> (46) 26931.78
> (47) 25153.13
> (48) 24904.98
> (49) 23888.79
> (50) 23311.97
> (51) 22239.48
> (52) 20958.73
> (53) 19483.88
> (54) 55.60012
> (55) 43.17431
> (56) 40.45679
> (57) 37.83539
> (58) 34.2077
> (59) 33.30375
> (60) 30.14031
> (61) 28.53269
> (62) 28.05641
> (63) 26.14609
> (64) 24.2809
> (65) 23.80275
> (66) 20.61175
> (67) 19.54397
> (68) 18.97411
> (69) 17.43671
> (70) 16.92381
> (71) 14.8029
> (72) 14
> (73) 13.2647
> (74) 12.197
> (75) 11.47018
> (76) 8.760074
> (77) 6.609896
> (78) 0
> (79) 0
> (80) 0
> (81) 0
> (82) 0
> (83) 0
> (84) 0
> (85) 0
> (86) 0
> .
> .
> .
> -----------------------------------------------------------------------------------
> The NCL script:
> ;*********************************************
> load "/cygdrive/E/cygwin/usr/local/lib/ncarg/nclscripts/csm/gsn_code.ncl"
> load "/cygdrive/E/cygwin/usr/local/lib/ncarg/nclscripts/csm/gsn_csm.ncl"
> load "/cygdrive/E/cygwin/usr/local/lib/ncarg/nclscripts/csm/contributed.ncl"
> load "/cygdrive/E/cygwin/usr/local/lib/ncarg/nclscripts/csm/shea_util.ncl"
> begin
> ;********************************************************
> filename1 = "/cygdrive/E/cygwin/home/stone/sst_filter_158point.asc"
> filename2 = "/cygdrive/E/cygwin/home/stone/JJA_53-07_156point.asc"
> ;**************************************************************
> data1 = asciiread(filename1,(/158,55/),"float")
> data2 = asciiread(filename2,(/156,55/),"float")
> sst = new((/158,55/),float)
> rain = new((/156,55/),float)
> sst(:,:) = data1(:,:)
> rain(:,:)= data2(:,:)
>
> print("apply CCA method..")
> opt=False
> canr=cancor(sst,rain,opt)
> print("calculate propability")
> ;prob=gammainc(0.5*canr@chisq, 0.5*canr@ndof)
>
> print("output CCA")
> print(canr)
> ;print(canr@ndof)
> ;print(prob)
> end
> --------------------------------------------------
> --------------------------------------------------------------------------------------
>
>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Apr 14 05:52:24 2010
This archive was generated by hypermail 2.1.8 : Wed Apr 14 2010 - 09:15:22 MDT