modelli

CICLO PER MODELLI FD

# Lista delle variabili per il ciclo
codes <- c("all_", "I_", "AV_I_", "PE_I_")

file_names <- c("FD_H", "FD_SLA", "FD_poll", "FD_seed_weight", "FD_life_span", "FD_N_fixing", "FD_growth_form")

# variabili_risposta <- c("canopy_height", "SLA", "pollinators", "seed_weight", "life_span", "N.fixing", "growth_form") 

elenco_modelli_FD<- c()

# Ciclo per fare i modelli in base ai sottogruppi (fila/interfila, AV/PE)
for (code in codes){
  # Ciclo per inserire i dataset
  for (file_name in file_names){
    
    dataframe_name <- paste0(code, file_name, sep="")
    dataframe <- get(dataframe_name)

  # Ciclo per costruire i modelli in base alla variabile risposta
    variabile_risposta <- colnames(dataframe)[1]
  modello <- lmer(as.formula(paste(variabile_risposta, "~ sistema + anno + (1 | blocco)")), data = dataframe)
  
  
  ## ANALISI MODELLO
  #risultati modello
  anova <- anova(modello)
  #shapiro test modello per normalità residui
  normality <- shapiro.test(residuals(modello))
  #Durbin-Watson test per indipendenza residui
  indipendence <- durbinWatsonTest(residuals(modello))

  
  # • SALVO OUTPUT MODELLI E ANALISI
  # Nome del modello basato sulla variabile risposta
  nome_modello <- paste(code, "FD_model_", file_name, sep = "")
  # Assegna il modello a una variabile con nome dinamico
  assign(nome_modello, modello)
  
  nome_anova <- paste(code, "FD_anova_model_", file_name, sep = "")
  # Assegna il modello a una variabile con nome dinamico
  assign(nome_anova, anova)
  
  nome_normality <- paste(code, "FD_normality_model_", file_name, sep = "")
  # Assegna il modello a una variabile con nome dinamico
  assign(nome_normality, normality$p.value)
  
  nome_indipendence <- paste(code, "FD_indipendence_model_", file_name, sep = "")
  # Assegna il modello a una variabile con nome dinamico
  assign(nome_indipendence, indipendence)
  
  
  
  # • Creo vettore con tutte le variabili create
  nome_modello <- paste(code, "FD_model_", file_name, sep = "")
  elenco_modelli_FD <- c(elenco_modelli_FD, nome_modello)
  elenco_modelli_FD <- unique(elenco_modelli_FD)
}}

## boundary (singular) fit: see help('isSingular')
## boundary (singular) fit: see help('isSingular')

r_squared <- numeric(length(elenco_modelli_CWM))
print(elenco_modelli_CWM)

##  [1] "all_CWM_model_canopy_height"    "all_CWM_model_SLA"             
##  [3] "all_CWM_model_seed_weight"      "all_CWM_model_N.fixing"        
##  [5] "all_CWM_model_annuals"          "all_CWM_model_perennials"      
##  [7] "all_CWM_model_chamaephyte"      "all_CWM_model_hemicryptophyte" 
##  [9] "all_CWM_model_cryptophyte"      "all_CWM_model_therophyte"      
## [11] "all_CWM_model_epiphyte"         "I_CWM_model_canopy_height"     
## [13] "I_CWM_model_SLA"                "I_CWM_model_seed_weight"       
## [15] "I_CWM_model_N.fixing"           "I_CWM_model_annuals"           
## [17] "I_CWM_model_perennials"         "I_CWM_model_chamaephyte"       
## [19] "I_CWM_model_hemicryptophyte"    "I_CWM_model_cryptophyte"       
## [21] "I_CWM_model_therophyte"         "I_CWM_model_epiphyte"          
## [23] "AV_I_CWM_model_canopy_height"   "AV_I_CWM_model_SLA"            
## [25] "AV_I_CWM_model_seed_weight"     "AV_I_CWM_model_N.fixing"       
## [27] "AV_I_CWM_model_annuals"         "AV_I_CWM_model_perennials"     
## [29] "AV_I_CWM_model_chamaephyte"     "AV_I_CWM_model_hemicryptophyte"
## [31] "AV_I_CWM_model_cryptophyte"     "AV_I_CWM_model_therophyte"     
## [33] "AV_I_CWM_model_epiphyte"        "PE_I_CWM_model_canopy_height"  
## [35] "PE_I_CWM_model_SLA"             "PE_I_CWM_model_seed_weight"    
## [37] "PE_I_CWM_model_N.fixing"        "PE_I_CWM_model_annuals"        
## [39] "PE_I_CWM_model_perennials"      "PE_I_CWM_model_chamaephyte"    
## [41] "PE_I_CWM_model_hemicryptophyte" "PE_I_CWM_model_cryptophyte"    
## [43] "PE_I_CWM_model_therophyte"      "PE_I_CWM_model_epiphyte"

for (model_name in elenco_modelli_CWM) {
  modello <- get(model_name)
  # summary_model <- summary(modello)
  R2 <- r.squaredGLMM(modello)
  index <- which(elenco_modelli_CWM == model_name)
  r_squared[index] <- R2
}

## Warning: 'r.squaredGLMM' now calculates a revised statistic. See the help page.

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

## Warning in r_squared[index] <- R2: number of items to replace is not a multiple
## of replacement length

print(r_squared)

##  [1] 0.0272093875 0.0319001247 0.0389749462 0.0109817163 0.0092817757
##  [6] 0.0092817757 0.0087425292 0.0616135215 0.1108975811 0.0211682382
## [11] 0.0515524711 0.0144521271 0.0161460010 0.0109288868 0.0118258133
## [16] 0.0724094136 0.0724094136 0.0206553116 0.0652549122 0.0297390626
## [21] 0.0731012990 0.0591131162 0.0189172442 0.0073341578 0.0005487211
## [26] 0.0076482851 0.0372676292 0.0372676292 0.0165344018 0.0269380861
## [31] 0.0395567068 0.0503034892 0.1181905652 0.0243029399 0.0738942794
## [36] 0.0312564790 0.0246485743 0.2970680771 0.2970680771 0.0330892126
## [41] 0.2000893439 0.0689777299 0.2741058790 0.0122147925

for (model_name in elenco_modelli_CWM) {
    i <- which(elenco_modelli_CWM == model_name)
    
  cat(model_name,": R quadro =", r_squared[i], "\n")
}

## all_CWM_model_canopy_height : R quadro = 0.02720939 
## all_CWM_model_SLA : R quadro = 0.03190012 
## all_CWM_model_seed_weight : R quadro = 0.03897495 
## all_CWM_model_N.fixing : R quadro = 0.01098172 
## all_CWM_model_annuals : R quadro = 0.009281776 
## all_CWM_model_perennials : R quadro = 0.009281776 
## all_CWM_model_chamaephyte : R quadro = 0.008742529 
## all_CWM_model_hemicryptophyte : R quadro = 0.06161352 
## all_CWM_model_cryptophyte : R quadro = 0.1108976 
## all_CWM_model_therophyte : R quadro = 0.02116824 
## all_CWM_model_epiphyte : R quadro = 0.05155247 
## I_CWM_model_canopy_height : R quadro = 0.01445213 
## I_CWM_model_SLA : R quadro = 0.016146 
## I_CWM_model_seed_weight : R quadro = 0.01092889 
## I_CWM_model_N.fixing : R quadro = 0.01182581 
## I_CWM_model_annuals : R quadro = 0.07240941 
## I_CWM_model_perennials : R quadro = 0.07240941 
## I_CWM_model_chamaephyte : R quadro = 0.02065531 
## I_CWM_model_hemicryptophyte : R quadro = 0.06525491 
## I_CWM_model_cryptophyte : R quadro = 0.02973906 
## I_CWM_model_therophyte : R quadro = 0.0731013 
## I_CWM_model_epiphyte : R quadro = 0.05911312 
## AV_I_CWM_model_canopy_height : R quadro = 0.01891724 
## AV_I_CWM_model_SLA : R quadro = 0.007334158 
## AV_I_CWM_model_seed_weight : R quadro = 0.0005487211 
## AV_I_CWM_model_N.fixing : R quadro = 0.007648285 
## AV_I_CWM_model_annuals : R quadro = 0.03726763 
## AV_I_CWM_model_perennials : R quadro = 0.03726763 
## AV_I_CWM_model_chamaephyte : R quadro = 0.0165344 
## AV_I_CWM_model_hemicryptophyte : R quadro = 0.02693809 
## AV_I_CWM_model_cryptophyte : R quadro = 0.03955671 
## AV_I_CWM_model_therophyte : R quadro = 0.05030349 
## AV_I_CWM_model_epiphyte : R quadro = 0.1181906 
## PE_I_CWM_model_canopy_height : R quadro = 0.02430294 
## PE_I_CWM_model_SLA : R quadro = 0.07389428 
## PE_I_CWM_model_seed_weight : R quadro = 0.03125648 
## PE_I_CWM_model_N.fixing : R quadro = 0.02464857 
## PE_I_CWM_model_annuals : R quadro = 0.2970681 
## PE_I_CWM_model_perennials : R quadro = 0.2970681 
## PE_I_CWM_model_chamaephyte : R quadro = 0.03308921 
## PE_I_CWM_model_hemicryptophyte : R quadro = 0.2000893 
## PE_I_CWM_model_cryptophyte : R quadro = 0.06897773 
## PE_I_CWM_model_therophyte : R quadro = 0.2741059 
## PE_I_CWM_model_epiphyte : R quadro = 0.01221479

r.squaredGLMM(PE_I_CWM_model_canopy_height)

##             R2m        R2c
## [1,] 0.02430294 0.06250206

r.squaredGLMM(AV_I_CWM_model_canopy_height)

##             R2m        R2c
## [1,] 0.01891724 0.02055278

r.squaredGLMM(I_CWM_model_canopy_height)

##             R2m        R2c
## [1,] 0.01445213 0.03277968

r.squaredGLMM(all_CWM_model_canopy_height)

##             R2m        R2c
## [1,] 0.02720939 0.08087853

miglior_modello <- which.max(r_squared)

cat("\nIl miglior modello è il numero", miglior_modello, "con R quadro =", r_squared[miglior_modello], "\n")

## 
## Il miglior modello è il numero 39 con R quadro = 0.2970681

# 
# 
# for (model_name in elenco_modelli_CWM) {
#   modello <- get(model_name)
#   summary_model <- summary(modello)
#   index <- which(elenco_modelli_CWM == model_name)
#   r_squared[index] <- summary_model$r.squared
# }



tab_model(modello)

	epiphyte
Predictors	Estimates	CI	p
(Intercept)	0.00	-0.00 – 0.00	0.199
sistema [ICC]	-0.00	-0.00 – 0.00	0.223
sistema [INC]	-0.00	-0.00 – -0.00	0.019
anno	0.00	-0.00 – 0.00	0.254
Random Effects
σ²	0.00
τ₀₀ _blocco	0.00
ICC	0.02
N _blocco	3
Observations	586
Marginal R² / Conditional R²	0.012 / 0.032

modelli

Andrea De Toma

2023-06-23

Ciclo per caricare i dataframe FD e CWM

Ciclo per costruire modelli CWM

CICLO PER MODELLI FD