-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathwindow.R
228 lines (184 loc) · 10.6 KB
/
window.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
#######Data input#########
dat <- read.table(file = "C:/Users/Walid/Desktop/Work/DataManp/Data/R-IAC_Data_230210-sub.txt", header = TRUE,stringsAsFactors = FALSE)
df <- cbind.data.frame("BIRTHY" = dat$BIRTHYW, "YEARM" = dat$YEARM, "Ind" = dat$FEMME, "AFR" = dat$AFR, "InfMort" = dat$OffMortality)
Temper <- read.table(file = "C:/Users/Walid/Desktop/Work/DataManp/Data/Temperature.txt", header = TRUE, stringsAsFactors = FALSE)
PDSI <- read.table(file = "C:/Users/Walid/Desktop/Work/DataManp/Data/PDSI_local.txt", header = TRUE, stringsAsFactors = FALSE)
NAO <- read.table(file = "C:/Users/Walid/Desktop/Work/DataManp/Data/NAO-winter-1400-2001.txt", header = TRUE, stringsAsFactors = FALSE)
Solar <- read.table(file = "C:/Users/Walid/Desktop/Work/DataManp/Data/Solar.txt", header = TRUE, stringsAsFactors = FALSE)
mortrate <- read.table(file = "C:/Users/walid/Desktop/Work/Mortality/mortrate.txt", header = TRUE)
dat$AFR <- as.numeric(scale(dat$AFR, center = TRUE, scale = TRUE))
dat$Temper<- as.numeric(scale(dat$Temper, center = TRUE, scale = TRUE))
dat$NAO <- as.numeric(scale(dat$NAO, center = TRUE, scale = TRUE))
dat$Mortrate <- as.numeric(scale(dat$Mortrate, center = TRUE, scale = TRUE))
dat$Solar <- as.numeric(scale(dat$Solar, center = TRUE, scale = TRUE))
dat$PDSI <- as.numeric(scale(dat$PDSI, center = TRUE, scale = TRUE))
colnames(NAO)[2] <- "value"
colnames(PDSI)[2] <- "value"
colnames(Solar)[2] <- "value"
colnames(PDSI)[1] <- "Year"
colnames(Solar)[1] <- "Year"
colnames(mortrate)[2] <- "value"
#########correlation coefficients######
corr.coef <- function(data, col1, col2){
#calculte the mean values of col1 annually
y <- aggregate(data[,col1], list("BIRTHY" = data[,1]), mean)
#calculte the mean values of col2 annually
x <- aggregate(data[,col2], list("BIRTHY" = data[,1]), mean)
colnames(y)[2] <- as.character(col1)
colnames(x)[2] <- as.character(col2)
temp.df <- merge(y,x, by = "BIRTHY") #merge the datasets for correlation calculation
coef.window <- cor(temp.df[,2], temp.df[,3], method = "pearson") #calculate Pearson's correlation coefficient
return(coef.window)
}
#######efficient way to calculte the correaltion coefficients of all windows####
columns <- names(test[,2:37]) #assign all the columns to be analyzed
coeff <- NULL #create empty dataframe
n <- 1 #allows succession for loop
for(col in columns){
coeff[n] <- corr.coef(test, "AFR", col) #calculte the correlation coefficient using the created function
n <- n + 1
}
write.table(coeff,file = "C:/Users/Walid/Desktop/coeff.txt") #outputs all the correlation coefficients for all the created windows
############group by temperature#######################
for(i in 1:nrow(test)){
test$tempgrp <- NULL
if(i %in% 1:143){
test$temgrp[i] <- as.character("group1")
}else if(i %in% 144:286){
test$temgrp[i] <- as.character("group2")
}else if(i %in% 287:429){
test$temgrp[i] <- as.character("group3")
}else if(i %in% 430:572){
test$temgrp[i] <- as.character("group4")
}
}
#########group by 1st and 3rd Quartile##########
dat$mortrate2<- with(dat, factor(findInterval(dat$mortrate_01y,
c(-Inf, quantile(dat$Mortrate, probs=c(0.25, .75)), Inf)),
labels=c("Low","Middle","High")))
library(Hmisc)
dat$mortrate2 <- cut2(dat$Mortrate, g =3)
##########################window function#########################
test = data.frame(cbind("BIRTHYW"=dat$BIRTHYW,column=rep(0,572)))
window <- function(distance=0, len=0,dataframe, variable){
for(m in distance){
for(y in len){
for(i in 1:nrow(dataframe)){
start <- dataframe$BIRTHY[i] + m #start year of window
end <- start + y #end year of window
years <- seq(start,end,by = 1) #create intra-years
z = seq(1, length(years),by = 1)
temp <- NULL
#extract the variable values for each year
for(n in z){
temp[n] <- variable[variable$Year == years[n],]$value
}
mean.temp <- mean(temp) #calculte the mean variable for the window
test$column[i] <<- mean.temp #assign it to proper ind into intial dataframe
}
#naming the variable
name <- paste(deparse(substitute(variable)), as.character(m),sep = "+" )
colnames(test)[dim(test)[2]] <<- paste(as.character(name), as.character(y), sep = "_")
}
}
}
window(distance = 0, len = seq(0,3,1), dataframe = dat, variable = mortrate)
###REGRESSION MODELS#######################
##########Infant mortality##################
m <- glm(OffMortality~1+offset(log(FERTILITY)), data = dat, family = "quasipoisson")
m1 <- glm(OffMortality~poly(AFR,3)+mortrate4+COHORTW+TWIN+offset(log(FERTILITY)), data = dat, family = "quasipoisson")
m2 <- glm(OffMortality~AFR*mortrate1+COHORTW+TWIN+offset(log(FERTILITY)), data = dat, family = "quasipoisson")
m2 <- glm(OffMortality~scale(AFR)*mortrate4+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+COHORTW+TWIN+offset(log(FERTILITY)), data = dat, family = "quasipoisson")
anova(m1, m2, test = "Chisq")
Anova(m2, type = "II")
summary(m)
summary(m1)
summary(m2)
library(effects)
plot(effect("poly(AFR, 3)", m1, xlevels = list(AFR = seq(18,35,1),FERTILITY=seq(1,10,2))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE, ylim=c(0,2))
plot(effect("poly(AFR, 3)*mortrate1", m2, xlevels = list(AFR=seq(18,35,1),FERTILITY=seq(1,10,2))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE, ylim = c(0,2))
plot(allEffects(m1, default.levels = 50), rescale.axis = FALSE, multiline = TRUE)
#########LRS########
mmm <- glm(LRS~1, data = temp, family = "quasipoisson")
mmm1 <- glm(LRS~AFR+Mortrate+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+scale(BIRTHYW), data = dat, family = "quasipoisson")
mmm2 <- glm(LRS~AFR*Mortrate+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+scale(BIRTHYW)+OffMortality, data = dat, family = "quasipoisson")
mmm3 <- glm(LRS~AFR*mortrate1+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+COHORTW, data = dat, family = "quasipoisson")
mmm4 <- glm(LRS~scale(AFR)*mortrate1+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+COHORTW, data = dat, family = "quasipoisson")
anova(mmm4,mmm3, test = "Chisq")
Anova(mmm3, type = "II")
plot(effect("scale(AFR):scale(Mortrate)", mmm2, xlevels = list(AFR = seq(18,35,1), Mortrate = c(0.02274367,0.08612651,0.1733076))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("AFR:Mortrate", mmm2, xlevels = list(AFR = seq(18,35,1),Mortrate = c(0.02274367,0.08612651,0.1733076))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("AFR:mortrate1", mmm3, xlevels = list(AFR = seq(18,35,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
#########FERTILITY########
mmm <- glm(FERTILITY~1, data = temp, family = "quasipoisson")
mmm1 <- glm(FERTILITY~AFR+Mortrate+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+scale(BIRTHYW), data = dat, family = "quasipoisson")
mmm2 <- glm(FERTILITY~scale(AFR)*mortrate1+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+COHORTW+scale(OffMortality), data = dat, family = "quasipoisson")
mmm3 <- glm(FERTILITY~AFR*mortrate4+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+COHORTW, data = dat, family = "quasipoisson")
mmm4 <- glm(FERTILITY~scale(AFR)*mortrate1+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+COHORTW, data = dat, family = "quasipoisson")
anova(mmm4,mmm3, test = "Chisq")
Anova(mmm4, type = "II")
plot(effect("scale(AFR):scale(Mortrate)", mmm2, xlevels = list(AFR = seq(18,35,1), Mortrate = c(0.02274367,0.08612651,0.1733076))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("scale(AFR):mortrate1", mmm2, xlevels = list(AFR = seq(18,35,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("AFR:mortrate4", mmm3, xlevels = list(AFR = seq(18,35,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
#####fitdem######
mm <- lm(fitdem~1, data = dat)
mm1 <- lm(fitdem~AFR*FERTILITY+FERTILITY*Mortrate+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+scale(BIRTHYW)*Mortrate, data = dat)
mm1 <- lm(scale(fitdem)~scale(AFR)*scale(FERTILITY)+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+scale(BIRTHYW), data = dat)
mm2 <- lm(fitdem~AFR*mortrate1+COEFPAR+I(COEFPAR*COEFPAR)+TWIN, data = dat)
mm3 <- lm(scale(fitdem)~scale(AFR)*mortrate1+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+COHORTW, data = dat)
library(car)
Anova(mm2, type = "II")
af <- anova(mm2)
afss <- af$"Sum Sq"
print(cbind(af,PctExp=afss/sum(afss)*100))
summary(mm)
summary(mm2)
summary(mm3)
anova(mm1,mm2, test = "Chisq")
plot(effect("scale(AFR):scale(Mortrate)", mm1, xlevels = list(AFR = seq(18,35,1), Mortrate = c(0.02274367,0.08612651,0.1733076))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("mortrate1:COHORTW", mm2, xlevels = list(AFR=seq(18,35,1), COHORTW=seq(1,8,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("scale(AFR):mortrate1", mm3, xlevels = list(AFR = seq(18,35,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
##AFR###
model.null <- lm(AFR~1, data = dat)
model <- lm(scale(AFR)~scale(Mortrate_3y)+scale(BIRTHYW)+scale(COEFPAR)+scale(I(COEFPAR*COEFPAR))+TWIN+OffMortality, data = dat)
model <- lm(AFR~Mortrate_3y+BIRTHYW+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+OffMortality, data = dat)
#without 2.5% extremes
model_cens <- lm(AFR~Mortrate_3y+BIRTHYW+COEFPAR+I(COEFPAR*COEFPAR)+TWIN+OffMortality, data = test)
model.1 <- lm(AFR~mortrate4+COHORTW+TWIN, data = dat)
model.2 <- lm(AFR~mortrate4*COHORTW+TWIN, data = dat)
model.3 <- lm(scale(AFR)~mortrate1+COHORTW+TWIN+scale(OffMortality), data = dat)
Anova(model.1, type = "II")
anova(model.1, model.2)
af <- anova(model.1)
afss <- af$"Sum Sq"
print(cbind(af,PctExp=afss/sum(afss)*100))
summary(model)
summary(model.2)
plot(effect("scale(Mortrate)", model), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("mortrate_equal", model.1, xlevels = list()), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
plot(effect("mortrate1*COHORTW", model.1, xlevels = list(COHORTW=seq(1,8,1))), rescale.axis = FALSE, multiline = TRUE, rug = TRUE)
######################################################
a <- list("Low","Middle", "High")
b <- list("AFR", "FERTILITY","fitdem", "OffMortality", "Mortrate_3y")
for (i in a) {
for (j in b) {
print(paste(round(mean(dat[dat$mortrate1 == i,j]),2), i, j))
print(paste(round(sd(dat[dat$mortrate1 == i,j]),2), i,j))
}
}
for (i in 1:nrow(IAC)){
if(IAC$COHORTW[i] %in% seq(1,4,1)){
IAC$period[i] <- 1
}else{
IAC$period[i] <- 2
}
}
##################################################
#reproductive lifespan
for (i in 1:length(row.names(dat))) {
dat$repspan[i] <- abs(dat$AFR[i] - dat$ALR[i])
}
#rate of reproduction
for (i in 1:length(row.names(IAC))) {
dat$reprate[i] <- dat$FERTILITY[i]/dat$repspan[i]
}