Revision 25a68ae3
Added by Benoit Parmentier over 12 years ago
- ID 25a68ae32a5fdc3ee3af1594e88f4743d617db37
| climate/research/oregon/interpolation/Linear_reg.R | ||
|---|---|---|
| 16 | 16 |
path<-"C:/Data/Benoit/NCEAS/window_Oregon_data" |
| 17 | 17 |
setwd(path) |
| 18 | 18 |
#infile2<-"dates_interpolation_03012012.txt" # list of 10 dates for the regression |
| 19 |
infile2<-"dates_interpolation_03032012.txt"
|
|
| 19 |
infile2<-"dates_interpolation_03052012.txt"
|
|
| 20 | 20 |
prop<-0.3 #Proportion of testing retained for validation |
| 21 | 21 |
out_prefix<-"_03042012_r1" |
| 22 | 22 |
|
| ... | ... | |
| 24 | 24 |
|
| 25 | 25 |
###Reading the station data and setting up for models' comparison |
| 26 | 26 |
ghcn<-read.csv(paste(path,"/",infile1, sep=""), header=TRUE) #The "paste" function concatenates the path and file name in common string. |
| 27 |
ghcn = transform(ghcn,Northness = cos(ASPECT)) #Adding a variable to the dataframe |
|
| 28 |
ghcn = transform(ghcn,Eastness = sin(ASPECT)) #adding variable to the dataframe. |
|
| 29 |
|
|
| 27 | 30 |
ghcn = transform(ghcn,Northness_w = sin(slope)*cos(ASPECT)) #Adding a variable to the dataframe |
| 28 | 31 |
ghcn = transform(ghcn,Eastness_w = sin(slope)*sin(ASPECT)) #adding variable to the dataframe. |
| 29 | 32 |
set.seed(100) |
| 30 | 33 |
dates <-readLines(paste(path,"/",infile2, sep="")) |
| 31 | 34 |
|
| 32 |
results <- matrix(1,length(dates),10) #This is a matrix containing the diagnostic measures from the GAM models. |
|
| 35 |
results <- matrix(1,length(dates),14) #This is a matrix containing the diagnostic measures from the GAM models. |
|
| 36 |
|
|
| 33 | 37 |
ghcn.subsets <-lapply(dates, function(d) subset(ghcn, date==d)) #this creates a list of 10 subsets data |
| 34 | 38 |
#note that compare to the previous version date_ column was changed to date |
| 35 | 39 |
|
| ... | ... | |
| 51 | 55 |
####Regression part 2: GAM models |
| 52 | 56 |
|
| 53 | 57 |
GAM_ANUSPLIN1<-gam(tmax~ s(lat) + s (lon) + s (ELEV_SRTM), data=data_s) |
| 54 |
GAM_PRISM1<-gam(tmax~ s(lat) + s (lon) + s (ELEV_SRTM) + s (ASPECT)+ s(DISTOC), data=data_s)
|
|
| 58 |
GAM_PRISM1<-gam(tmax~ s(lat) + s (lon) + s (ELEV_SRTM) + s (Northness)+ s (Eastness) + s(DISTOC), data=data_s)
|
|
| 55 | 59 |
GAM_PRISM2<-gam(tmax~ s(lat) + s (lon) + s (ELEV_SRTM) + s (Northness_w)+ s (Eastness_w) + s(DISTOC), data=data_s) |
| 56 | 60 |
|
| 57 | 61 |
|
| ... | ... | |
| 60 | 64 |
results[i,1]<- dates[i] #storing the interpolation dates in the first column |
| 61 | 65 |
results[i,2]<- ns #number of stations used in the training stage |
| 62 | 66 |
|
| 63 |
results[i,5]<- AIC (GAM_ANUSPLIN1) |
|
| 64 |
results[i,6]<- AIC (GAM_PRISM1) |
|
| 65 |
GCVA1<-GAM_ANUSPLIN1$gcv.ubre |
|
| 66 |
results[i,7]<- GCVA1 |
|
| 67 |
GCVP1<-GAM_PRISM1$gcv.ubre |
|
| 68 |
results[i,8]<- GCVP1 |
|
| 67 |
results[i,6]<- AIC (GAM_ANUSPLIN1) |
|
| 68 |
results[i,7]<- AIC (GAM_PRISM1) |
|
| 69 |
results[i,8]<- AIC (GAM_PRISM2) |
|
| 70 |
|
|
| 71 |
results[i,9]<- GAM_ANUSPLIN1$gcv.ubre |
|
| 72 |
results[i,10]<- GAM_PRISM1$gcv.ubre |
|
| 73 |
results[i,11]<- GAM_PRISM2$gcv.ubre |
|
| 69 | 74 |
|
| 70 |
results[i,9]<-GAM_ANUSPLIN1$deviance |
|
| 71 |
results[i,10]<-GAM_PRISM1$deviance |
|
| 75 |
results[i,12]<-GAM_ANUSPLIN1$deviance |
|
| 76 |
results[i,13]<-GAM_PRISM1$deviance |
|
| 77 |
results[i,14]<-GAM_PRISM2$deviance |
|
| 72 | 78 |
|
| 73 | 79 |
#####VALIDATION: Prediction checking the results using the testing data######## |
| 74 | 80 |
|
| 75 | 81 |
y_pgANUSPLIN1<- predict(GAM_ANUSPLIN1, newdata=data_v, se.fit = TRUE) #Using the coeff to predict new values. |
| 76 | 82 |
y_pgPRISM1<- predict(GAM_PRISM1, newdata=data_v, se.fit = TRUE) |
| 77 |
|
|
| 83 |
y_pgPRISM2<- predict(GAM_PRISM2, newdata=data_v, se.fit = TRUE) |
|
| 84 |
|
|
| 78 | 85 |
res_ypgA1<- data_v$tmax - y_pgANUSPLIN1$fit #Residuals for GMA model that resembles the ANUSPLIN interpolation |
| 79 | 86 |
res_ypgP1<- data_v$tmax - y_pgPRISM1$fit #Residuals for GAM model that resembles the PRISM interpolation |
| 80 |
|
|
| 87 |
res_ypgP2<- data_v$tmax - y_pgPRISM2$fit |
|
| 88 |
|
|
| 81 | 89 |
RMSE_ypgA1 <- sqrt(sum(res_ypgA1^2)/nv) |
| 82 | 90 |
RMSE_ypgP1 <- sqrt(sum(res_ypgP1^2)/nv) |
| 91 |
RMSE_ypgP2 <- sqrt(sum(res_ypgP2^2)/nv) |
|
| 83 | 92 |
|
| 84 | 93 |
results[i,3]<-RMSE_ypgA1 |
| 85 | 94 |
results[i,4]<-RMSE_ypgP1 |
| 95 |
results[i,5]<-RMSE_ypgP2 |
|
| 96 |
|
|
| 97 |
data_name<-paste("ghcn_v_",dates[[i]],sep="")
|
|
| 98 |
assign(data_name,data_v) |
|
| 99 |
#ghcn_v<-ls(pattern="ghcn_v_") |
|
| 86 | 100 |
|
| 87 | 101 |
# end of the for loop #1 |
| 88 | 102 |
} |
| ... | ... | |
| 95 | 109 |
# Now turn it into a data.frame... |
| 96 | 110 |
|
| 97 | 111 |
results_table<-as.data.frame(results_num) |
| 98 |
colnames(results_table)<-c("dates","ns","RMSE_A1", "RMSE_P1", "AIC_A1", "AIC_P1", "GCV_A1", "GCV_P1", "Deviance_A1", "Deviance_P1")
|
|
| 112 |
colnames(results_table)<-c("dates","ns","RMSE_A1", "RMSE_P1","RMSE_P2", "AIC_A1", "AIC_P1", "AIC_P2", "GCV_A1", "GCV_P1", "GCV_P2", "Deviance_A1", "Deviance_P1", "Deviance_P2")
|
|
| 99 | 113 |
|
| 100 | 114 |
win.graph() |
| 101 |
barplot(results_table$RMSE_A1,main="RMSE for the A1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date")
|
|
| 115 |
barplot(results_table$RMSE_A1/10,main="RMSE for the A1 models",names.arg=results_table$dates,ylab="Temp (deg. C)",xlab="interolated date")
|
|
| 102 | 116 |
savePlot(paste("GAM_ANUSPLIN1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
| 103 | 117 |
win.graph() |
| 104 |
barplot(results_table$RMSE_P1,main="RMSE for the P1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 118 |
barplot(results_table$RMSE_P1/10,main="RMSE for the P1 models",names.arg=results_table$dates,ylab="Temp ( deg. C)",xlab="interolated date") |
|
| 119 |
savePlot(paste("GAM_PRISM1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
|
| 120 |
win.graph() |
|
| 121 |
barplot(results_table$RMSE_P2/10,main="RMSE for the P2 models",names.arg=results_table$dates,ylab="Temp ( deg. C)",xlab="interolated date") |
|
| 105 | 122 |
savePlot(paste("GAM_PRISM1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
| 106 | 123 |
win.graph() |
| 107 |
barplot(results_table$AIC_P1,main="AIC for the P1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date")
|
|
| 124 |
barplot(results_table$AIC_A1,main="AIC for the A1 models",names.arg=results_table$dates,ylab="Temp ( deg. C)",xlab="interolated date")
|
|
| 108 | 125 |
savePlot(paste("GAM_PRISM1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
| 109 | 126 |
win.graph() |
| 110 |
barplot(results_table$AIC_A1,main="AIC for the A1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date")
|
|
| 127 |
barplot(results_table$AIC_P1/10,main="AIC for the P1 models",names.arg=results_table$dates,ylab="Temp ( deg. C)",xlab="interolated date")
|
|
| 111 | 128 |
savePlot(paste("GAM_PRISM1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
| 112 | 129 |
win.graph() |
| 113 |
barplot(results_table$Deviance_A1,main="Deviance for the A1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 130 |
barplot(results_table$AIC_P2/10,main="AIC for the P2 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 131 |
savePlot(paste("GAM_PRISM1_RMSE",out_prefix,".emf", sep=""), type="emf")
|
|
| 132 |
win.graph() |
|
| 133 |
barplot(results_table$Deviance_A1/10,main="Deviance for the A1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 114 | 134 |
savePlot(paste("GAM_ANUSPLIN1_Deviance",out_prefix,".emf", sep=""), type="emf")
|
| 115 | 135 |
win.graph() |
| 116 |
barplot(results_table$Deviance_P1,main="Deviance for the P1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 136 |
barplot(results_table$Deviance_P1/10,main="Deviance for the P1 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date")
|
|
| 117 | 137 |
savePlot(paste("GAM_PRISM1_Deviance",out_prefix,".emf", sep=""), type="emf")
|
| 138 |
win.graph() |
|
| 139 |
barplot(results_table$Deviance_P2/10,main="Deviance for the P2 models",names.arg=results_table$dates,ylab="Temp (0.1 X deg. C)",xlab="interolated date") |
|
| 140 |
savePlot(paste("GAM_PRISM2_Deviance",out_prefix,".emf", sep=""), type="emf")
|
|
| 141 |
|
|
| 118 | 142 |
write.csv(results_table, file= paste(path,"/","results_GAM_Assessment",out_prefix,".txt",sep="")) |
| 119 | 143 |
|
| 120 | 144 |
|
| ... | ... | |
| 142 | 166 |
# #vis.gam(GAM_ANUSPLIN1, view=c("lat","ELEV_SRTM", theta=100,phi=200))
|
| 143 | 167 |
# savePlot("GAM_ANUSPLIN1_prediction2.emf", type="emf")
|
| 144 | 168 |
# |
| 145 |
# #results_val <-c(res_yplA1,res_ypgA1,res_yplP1,res_ypgP1) |
|
| 146 |
# results_val<-data.frame(res_yplA1=res_yplA1,res_ypgA1=res_ypgA1,res_yplP1=res_yplP1,res_ypgP1=res_ypgP1) |
|
| 147 |
# nv<- nrow(ghcn1507_v) |
|
| 148 |
# |
|
| 149 |
# mod_name<-c("yplA1","ypgA1","yplP1","ypgP1")
|
|
| 150 |
# mod_type<-c("lm_ANUSPLIN1","GAM_ANUSPLIN1","lm_PRISM1","GAM_PRISM1")
|
|
| 151 |
# |
|
| 152 |
# RMSE_all<-c(RMSE_yplA1,RMSE_ypgA1,RMSE_yplP1,RMSE_ypgP1) |
|
| 153 |
# AIC_all<-AIC(lm_ANUSPLIN1,GAM_ANUSPLIN1,lm_PRISM1,GAM_PRISM1) |
|
| 154 |
# GCV_all<-c(0,GCVA1,0,GCVP1) #This places the GCV values for each model in a vector |
|
| 155 |
# Deviance_all<-c(0,DevianceA1,0,DevianceP1) |
|
| 156 |
# |
|
| 157 |
# #results<-data.frame(model=mod_name,RMSE=RMSE_all,df=AIC_all$df,AIC=AIC_all$AIC) |
|
| 158 |
# #results_val<-data.frame(model=mod_name,RMSE=RMSE_all,df=AIC_all$df,AIC=AIC_all$AIC) |
|
| 159 |
# results<-data.frame(model=mod_name,RMSE=RMSE_all,df=AIC_all$df,AIC=AIC_all$AIC, GCV=GCV_all,Deviance=Deviance_all) |
|
| 160 |
# |
|
| 161 |
# #Add deviance |
|
| 162 |
# |
|
| 163 |
# barplot(results$RMSE,main="RMSE for the models",names.arg=c("yplA1","ypgA1","yplP1","ypgP1"),ylab="Temp (deg. C)")
|
|
| 164 |
# barplot(results$AIC,main="AIC for the models",names.arg=c("yplA1","ypgA1","yplP1","ypgP1"),ylab="AIC")
|
|
| 165 |
# #dump("results", file= paste(path,"/","results_reg1507.txt",sep=""))
|
|
| 166 |
# write.csv(results, file= paste(path,"/","results_reg1507_Assessment.txt",sep="")) |
|
| 167 |
# #write.csv(results_val, file= paste(path,"/","results_reg1507_val.txt",sep="")) |
|
| 169 |
|
|
| 168 | 170 |
# |
| 169 | 171 |
|
| 170 | 172 |
|
Also available in: Unified diff
Transform aspect in Eastness and Northness variables in the GAM.