Iuse mixture discriminant analysis to do a binary classify.
HereI have some question about it. I think it's just the basic R knowledge. I hopesomeone could help me.
Hereis my R-code it's simple.
library(mda)
mda.fit=mda(Result~Cu+Fe+Si+Zn+Mn+Mg+CY+IP+BL+IRT+FT+IS+PS,data=data,iter=10)
summary(mda.fit)
mda.fit$fit$coef
coef(mda.fit)
Ihave two question about coefficients and one about iteration.
1I want know why here the coefficients has 5 columns, and what's the meaning foreach column?
2Why the result i use coef(mda.fit) is different with the result i get from theresult using mda.fit$fit$coef
mda.fit$fit$coef
[,1][,2] [,3] [,4] [,5]
Intercept -2.466366e+01 -4.995704e-01 18.0846025065 27.0392560847 -4.493157e+01
Cu 2.230663e+00 8.906730e-01 -1.3142124517 -1.7159225257 3.131520e+00
Fe 2.516078e+00 -1.551403e+00 -3.3118497077 -3.2039223892 5.809446e+00
Si 1.337944e-02 8.899690e-03 0.0074259910 0.0023786400 9.839065e-03
Zn -9.301386e+00 8.589747e+00 6.9499591909 5.6767075645 -1.518349e+01
Mn 6.130803e-02 -4.224064e-01 -0.1840648468 -0.2755888061 3.502979e-01
Mg 6.750485e-01 -8.582770e-01 -0.2688717403 -0.2940197351 6.773331e-01
CY 4.366998e-03 1.865515e-04 -0.0027944267 -0.0035768707 -4.397858e-03
IP 1.061739e-04 -2.925672e-06 -0.0005207723 -0.0000901363 1.101494e-04
BL -4.101475e-03 8.702419e-03 0.0009864619 0.0041132666 -7.836253e-04
IRT 9.291208e-03 2.069711e-04 -0.0130543798 -0.0037143474 6.371939e-03
FT -6.207071e+00 3.549535e+00 6.5869820877 7.1207680797 -1.854928e+01
IS 1.657820e-02 4.666378e-04 -0.0064468501 -0.0150353407 2.946918e-02
PS 3.215113e+00 1.511803e-02 -2.3666182682 -3.5782237326 6.102600e+00
> coef(mda.fit)
[,1] [,2] [,3] [,4] [,5]
Intercept -6.446274e+01 -1.086662e+02 -1.767562e+01 6.402725e+00 5.359666e+00
Cu 4.139227e+00 8.027177e+00 1.193511e+01 1.548998e+00 -2.328874e+01
Fe 8.634991e+00 1.374100e+01 -1.984376e+01 1.706935e+01 2.219714e+01
Si 1.589762e-02 1.401047e-02 1.217354e-01 -1.376659e-01 -6.080091e-01
Zn -2.272591e+01 -3.461331e+01 1.110304e+02 -1.742704e+01 1.157421e+02
Mn 5.746246e-01 8.052259e-01 -5.314858e+00 -2.258044e-01 7.262064e+00
Mg 8.630686e-01 1.868588e+00 -1.101212e+01 -2.292818e+00 -1.179283e+01
CY -1.526283e-02 8.490945e-03 5.559201e-04 -6.317106e-04 3.062405e-04
IP -6.651311e-05 7.501312e-04 -5.727482e-04 7.534739e-03 2.601896e-03
BL 2.614969e-03 -1.109719e-02 1.096480e-01 5.162840e-02 -1.159187e-02
IRT 1.167383e-03 3.100366e-02 -9.457535e-03 1.516631e-01 -1.185904e-01
FT -3.154272e+01 -3.472714e+01 4.310293e+01 -2.358301e+01 -5.394543e+00
IS 4.518206e-02 6.405354e-02 1.736533e-02 -7.389142e-02 -1.367332e-01
PS 8.923666e+00 1.441429e+01 1.763665e+00 -1.166375e+00 2.607331e-01
2I want to know how to decide the number of iteration. With different iterationnumber the result is different.
Ineed find a relation between the factors and response variable, not just find asmallest mse.
Thanks,for help.