I'm trying to make a PLS regressionon the following data:
indipendent variables
4.06 4.51 4.6 1.5 74.04 20.86 16.09 13.91 30.83
15.83 0.77 0.51 4.21 5.06 35.99 41.62 41.4 47.46
0.12 0.12 5.47 9.92 1.09 0.3
4.1 4.37 3.5 2.8 59.37 43.34 16.96 15.48 27.31
26.02 2.96 2.83 6.96 5.7 14.37 9.67 25.73 20.98
0.48 0.59 2.35 2.01 0.5 0.49
4.52 4.21 2.2 0.8 32.14 13.29 14.61 16.61 17.17
12.06 1.87 0.83 5.99 4.09 32.15 17.23 40.99 22.62
0.19 0.24 3.2 4.93 0.37 0.2
4.77 6.03 4.9 2.8 75.02 34.78 15.31 12.42 41.69
33.87 0.29 0.24 4.89 4.28 65.17 84.66 70.51 89.3
0.08 0.05 16.86 17.83 1 0.65
3.68 4.02 4.8 1.8 74.95 24.06 15.61 13.37 31.27
19.42 0.45 0.13 1.36 1.06 5.75 3.04 7.83 4.18
0.24 0.35 3.02 8.15 3.53 1.7
3.06 3.31 4.7 1.7 83.84 32.8 17.84 19.29 33.22
28.46 0.5 0.28 1.86 1.245 7.03 5.97 9.89 7.76
0.26 0.21 3.72 4.45 2.53 1.37
7.1 7.18 6 4.7 99.1 59.06 16.52 12.57 55.15
41.9 0.235 0.26 24.43 21.48 79.69 85.39 104.48 108.21
0.31 0.25 103.96 82.62 0.25 0.22
6.86 7.13 9.8 5.8 169.84 96.58 17.33 16.65 93.13
69.94 0.13 0.1 24.14 19.41 102.23 82.02 126.57 101.58
0.24 0.24 185.69 194.1 0.41 0.3
7.23 7.28 6.9 7 84.96 85.58 12.31 12.23 60.39
59.33 0.18 0.175 21.24 20.37 88.05 85.27 109.29 105.98
0.24 0.24 118 116.4 0.32 0.34
7.22 7.27 7.9 7.1 97.16 72.68 12.3 10.24 57.3
50.32 0.16 0.13 20.78 21.42 83.11 75.02 106.63 96.43
0.25 0.29 129.88 164.77 0.38 0.33
7.17 7.37 6 1.2 88.44 19.98 14.74 16.65 52.29
14.32 0.98 2.66 24.68 40.38 87.52 92.58 116.1 136.96
0.28 0.44 25.18 15.18 0.24 0.03
7.28 7.32 2.8 2.4 30.66 34.08 10.95 14.2 28.21
25.53 0.25 0.22 22.44 24 67.97 71.95 91 96.27
0.33 0.33 89.76 109.09 0.12 0.1
7.23 7.41 2 1 43.96 23.53 21.98 23.53 16.41
8.65 1.28 1.45 17.42 19.96 94.35 110.38 112.4 132.09
0.18 0.18 13.61 13.77 0.11 0.05
7.09 7.3 4 2.7 74.62 57.76 18.65 21.39 30.24
19.83 0.95 1.3 25.45 32.4 88.81 95.63 118.34 131.88
0.29 0.34 26.79 24.92 0.16 0.08
4.7 4.55 5.4 3.4 82.31 48.48 15.24 14.26 35.85
25.55 4.69 3.37 7.39 6.64 37.59 34.83 51.85 46.52
0.2 0.19 1.58 1.97 0.73 0.51
6.89 7.36 8.8 2 131.52 32.8 14.95 16.4 57.87
14.21 0.94 0.6 9.22 7.97 103.82 122.26 115.57 131.09
0.09 0.07 9.81 13.28 0.95 0.25
6.68 7.52 7.8 2.8 86.99 35.5 11.15 12.68 44.83
17.8 0.4 0.34 2.31 2.24 92 107.95 94.98 111.11
0.03 0.02 5.78 6.59 3.38 1.25
The variables are:
pH2 pH3 N2 N3 C org 2 C org 3 C/N2 C/N3 CSC2
CSC3 K2 K3 Mg2 Mg3 Ca2 Ca3 GSB2 GSB3
Mg/Ca2 Mg/Ca3 Mg/K2 Mg/K3 N/Mg2 N/Mg3
Dependent variable would be:
Trasparence of vegetation cover %
7.500
22.940
19.500
22.170
8.570
29.830
22.330
17.670
13.450
23.830
51.210
18.170
36.170
55.000
24.830
30.000
7.670
I linked these two table but when I ask to LineaReg to perform its
PLS->Modelling option with 1 or 2 components it says "Non positive
Variance".
Ehm, what does it mean?
If I use PlsGen2, instead, with 1 or 2 component, it gives me the message
"Unhandled exception: c0000005
At address: 00407af2"
What I'm doing wrongly ?
(Some time ago I noticed that PlsGen2 crashed on the same data on which
LineaReg|PLS->Modelling didn't. Next I loose the data so it was impossible
to report this fact that, to me, seemed a bug).
Thank you.
-------------------------------------------------------------------------------
Federico Spinazzi
email: federico@syspr03.disat.unimi.it
-----------------------------------------?-------------------------------------
This archive was generated by hypermail 2b30 : Sat Feb 10 2001 - 10:21:48 MET