ich habe mit dem lme4 package ein gemixtes Modell erstellt.
FA= numerisch (Flugaktivität)
treatment=faktoriell
To=numerisch (Außentemperatur)
week=integer (ID der Untersuchungswoche)
bunker=faktoriell (8 Bunker)
- Code: Alles auswählen
m1<-glmer(FA~treatment * To +week +(1|bunker) , data=model,family="poisson", na.action=na.exclude)
Bis hier hin ist auch alles gut. Wenn ich aber die Interaktion (To:week) einfüge, dann bekomme ich folgende Fehlermeldung
- Code: Alles auswählen
Warning messages:
1: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model failed to converge with max|grad| = 0.00144043 (tol = 0.001, component 1)
2: In checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
Model is nearly unidentifiable: very large eigenvalue
- Rescale variables?
Verstehe aber nicht warum. Wie sollte ich die Variablen anders skalieren?
Der Modeloutput:
- Code: Alles auswählen
Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
Family: poisson ( log )
Formula: FA ~ treatment * To + week + (To:week) + (1 | bunker)
Data: model
AIC BIC logLik deviance df.resid
1363.8 1384.4 -674.9 1349.8 133
Scaled residuals:
Min 1Q Median 3Q Max
-3.7898 -1.7585 -0.3913 1.2428 11.1604
Random effects:
Groups Name Variance Std.Dev.
bunker (Intercept) 0.07295 0.2701
Number of obs: 140, groups: bunker, 8
Fixed effects:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 1.7375000 0.1269460 13.687 < 2e-16 ***
treatmentmon 0.4457489 0.0589911 7.556 4.15e-14 ***
To 0.1095392 0.0194596 5.629 1.81e-08 ***
week 0.0444788 0.0123033 3.615 0.0003 ***
treatmentmon:To -0.0766273 0.0122636 -6.248 4.15e-10 ***
To:week -0.0006522 0.0026561 -0.246 0.8060
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Correlation of Fixed Effects:
(Intr) trtmnt To week trtm:T
treatmentmn -0.317
To -0.415 0.263
week -0.540 0.037 0.534
tretmntmn:T 0.126 -0.357 -0.325 0.014
To:week 0.373 -0.091 -0.876 -0.604 -0.050
convergence code: 0
Model failed to converge with max|grad| = 0.00144043 (tol = 0.001, component 1)
Model is nearly unidentifiable: very large eigenvalue
- Rescale variables?
Die Daten
- Code: Alles auswählen
structure(list(bunker = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 6L, 6L, 6L,
6L, 7L, 7L, 7L, 7L, 8L, 8L, 8L, 8L, 1L, 1L, 1L, 1L, 2L, 2L, 2L,
2L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 5L, 5L, 5L, 5L, 6L, 6L, 6L,
6L, 7L, 7L, 7L, 7L, 8L, 8L, 8L, 8L, 1L, 1L, 1L, 2L, 2L, 2L, 3L,
3L, 3L, 4L, 4L, 4L, 5L, 6L, 6L, 6L, 7L, 7L, 7L, 8L, 8L, 8L, 1L,
1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L, 4L, 4L, 4L, 5L, 6L, 6L, 6L, 7L,
7L, 7L, 8L, 8L, 8L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L,
6L, 7L, 7L, 8L, 8L, 1L, 1L, 2L, 2L, 3L, 3L, 4L, 4L, 5L, 5L, 6L,
6L, 7L, 7L, 8L, 8L), .Label = c("1", "2", "3", "4", "5", "6",
"7", "8"), class = "factor"), FA = c(2, 3, 7, 33, 22, 0, 1, 19,
10, 7, 6, 4, 5, 7, 8, 4, 23, 7, 4, 23, 25, 1, 1, 3, 8, 7, 2,
0, 8, 1, 7, 11, 12, 1, 16, 13, 7, 27, 20, 23, 27, 11, 12, 0,
15, 17, 10, 6, 14, 9, 11, 0, 9, 10, 4, 0, 3, 0, 5, 8, 4, 38,
18, 32, 18, 0, 2, 2, 14, 10, 15, 5, 7, 10, 14, 0, 1, 4, 5, 1,
0, 2, 10, 7, 12, 4, 3, 4, 0, 14, 6, 6, 15, 19, 12, 0, 3, 7, 5,
5, 3, 6, 5, 3, 19, 9, 18, 5, 0, 2, 10, 27, 17, 6, 4, 11, 10,
29, 0, 12, 1, 14, 1, 37, 18, 29, 8, 41, 3, 24, 14, 10, 7, 20,
44, 11, 9, 9, 15, 8), treatment = structure(c(1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("ctrl", "mon"), class = "factor"),
To = c(5, 0, 0, 7, 4.5, -1.5, -1, 5, 5, 0, 0, 7, 4.5, -1.5,
-1, 5, 5, 0, 0, 7, 4.5, -1.5, -1, 5, 5, 0, 0, 7, 4.5, -1.5,
-1, 5, 4.5, -1.5, -1, 5, 3, -2, 0, 6, 4.5, -1.5, -1, 5, 3,
-2, 0, 6, 4.5, -1.5, -1, 5, 3, -2, 0, 6, 4.5, -1.5, -1, 5,
3, -2, 0, 6, 5.5, -1.5, -1, -4.4, -1.1, -3, 5.5, -1.5, -1,
-4.4, -1.1, -3, -1, -4.4, -1.1, -3, 5.5, -1.5, -1, -4.4,
-1.1, -3, -4.4, -1.1, -3, -4.7, 2.2, -2.3, -4.4, -1.1, -3,
-4.7, 2.2, -2.3, -3, -4.7, 2.2, -2.3, -4.4, -1.1, -3, -4.7,
2.2, -2.3, -9.8, 2.7, -7.6, 7.6, -9.8, 2.7, -7.6, 7.6, -9.8,
2.7, -7.6, 7.6, -9.8, 2.7, -7.6, 7.6, -7.6, 7.6, -4, 6.4,
-7.6, 7.6, -4, 6.4, -7.6, 7.6, -4, 6.4, -7.6, 7.6, -4, 6.4
), week = c(1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L,
1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L,
4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L,
3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L, 2L, 3L, 4L, 1L,
2L, 3L, 4L, 1L, 2L, 3L, 4L, 5L, 6L, 7L, 5L, 6L, 7L, 5L, 6L,
7L, 5L, 6L, 7L, 7L, 5L, 6L, 7L, 5L, 6L, 7L, 5L, 6L, 7L, 5L,
6L, 7L, 5L, 6L, 7L, 5L, 6L, 7L, 5L, 6L, 7L, 7L, 5L, 6L, 7L,
5L, 6L, 7L, 5L, 6L, 7L, 8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L, 8L,
9L, 8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L,
8L, 9L, 8L, 9L, 8L, 9L, 8L, 9L)), row.names = c(NA, -140L
), class = "data.frame")
Ich freue mich über jeden Hinweis
Danke
