wilcox.test(hrp$TGHDLrate~hrp$type)
wilcox.test(hrp$TGGLUrate~hrp$type)
wilcox.test(hrp$TnI~hrp$type)
wilcox.test(hrp$myoglobin~hrp$type)
wilcox.test(hrp$CKMB~hrp$type)
wilcox.test(hrp$CK~hrp$type)
wilcox.test(hrp$cTnT~hrp$type)
wilcox.test(hrp$serumamylase~hrp$type)
wilcox.test(hrp$PH~hrp$type)
wilcox.test(hrp$PCO2~hrp$type)
wilcox.test(hrp$PO2~hrp$type)
wilcox.test(hrp$SO2~hrp$type)
wilcox.test(hrp$AB~hrp$type)
wilcox.test(hrp$SB~hrp$type)
wilcox.test(hrp$CO2~hrp$type)
wilcox.test(hrp$SBE~hrp$type)
wilcox.test(hrp$ABE~hrp$type)
wilcox.test(hrp$lacticacid~hrp$type)
wilcox.test(hrp$ionizedcalcium~hrp$type)
wilcox.test(hrp$bloodglucose~hrp$type)
wilcox.test(hrp$glycated.hemoglobin~hrp$type)
mytable <- xtabs(~UGLU+type,data = hrp)
mytable
chisq.test(mytable)
fisher.test(mytable)
mytable <- xtabs(~urineketone+type,data = hrp)
mytable
chisq.test(mytable)
mytable <- xtabs(~urineprotein+type,data = hrp)
mytable
chisq.test(mytable)
fisher.test(mytable)
wilcox.test(hrp$NLR~hrp$type)
wilcox.test(hrp$PLR~hrp$type)
wilcox.test(hrp$MLR~hrp$type)
wilcox.test(hrp$RPR~hrp$type)
wilcox.test(hrp$CLR~hrp$type)
wilcox.test(hrp$CAR~hrp$type)
is.data.frame(hrp)
install.packages("glmnet")
library(glmnet)
hrpcolnames <- colnames(hrp)
hrpcolnames
str(hrpcolnames)
hrpx <- hrpcolnames[15,18,33,35,39:42,44,47,49,54,58,59,61,62,63,66,67,69,71,72,74:76,78,81,83,84,96,97,99,101,102,109,112,114,115,118:125,129,137:139,141,142]
hrpx <- hrpcolnames[15,18,33,35,39,40,41,42,44,47,49,54,58,59,61,62,63,66,67,69,71,72,74,75,76,78,81,83,84,96,97,99,101,102,109,112,114,115,118,119,120,121,122,123,124,125,129,137,138,139,141,142]
names(hrp)
hrpx <- c("diabetes","replase","RBC","hct","RDW","WBC","neutrophilicgranulocyte.percentage", "Lymphocytepercentage","Percentage.of.eosinophils","Lymphocyte","eosinophils","PCT","CRP","PCT2","PT","PTA","INR","TT","Ddimer","Na","Ca","PO4","Mg","HCO3","BUN","UricAcid","cystatinC","AST","LDH","GGT","totalcholesterol","TG","LDL","ApolipoproteinB","ApolipoproteinA1Brate","CKMB","serumamylase","PH","PCO2","AB","SB","CO2","SBE","ABE","lacticacid","ionizedcalcium","bloodglucose","urineprotein","NLR","PLR","MLR","CLR","CAR")
hrp.lasso <- glmnet(x=hrpx,y=hrp$type,alpha = 1,family = "binomial",nlambda = 1000)
hrpX <- as.matrix(hrp[15,18,33,35,39,40,41,42,44,47,49,54,58,59,61,62,63,66,67,69,71,72,74,75,76,78,81,83,84,96,97,99,101,102,109,112,114,115,118,119,120,121,122,123,124,125,129,137,138,139,141,142])
hrpX <- as.matrix(hrp[,c(15,18,33,35,39,40,41,42,44,47,49,54,58,59,61,62,63,66,67,69,71,72,74,75,76,78,81,83,84,96,97,99,101,102,109,112,114,115,118,119,120,121,122,123,124,125,129,137,138,139,141,142)])
hrpY <- as.matrix(hrp[,136])
hrp.lasso <- glmnet(x=hrpX,y=hrpY,alpha = 1,family = "binomial",nlambda = 1000)
?makeX
hrpcomplete <- na.omit(hrp)
summary(hrpcomplete)
str(hrpcomplete)
hrpX1 <- makeX(hrp[,c(15,18,33,35,39,40,41,42,44,47,49,54,58,59,61,62,63,66,67,69,71,72,74,75,76,78,81,83,84,96,97,99,101,102,109,112,114,115,118,119,120,121,122,123,124,125,129,137,138,139,141,142)],na.impute = TRUE)
hrp.lasso <- glmnet(x=hrpX1,y=hrpY,alpha = 1,family = "binomial",nlambda = 1000)
plot(hrp.lasso)
plot(hrp.lasso,xvar = "lambda")
print(hrp.lasso)
hrp.lasso.coef <- predict(hrp.lasso,s=0.053,type = "coefficients")
hrp.lasso.coef
hrplasso.y <- predict(hrp.lasso,newx = hrpX1,type = "response",s=0.053)
plot(hrplasso.y,hrpY)
mytable  <-  xtabs(~hrplasso.y+hrpY)
mytable
hrplasso.y
hrplogisticfit <- glm(type~Ddimer+Ca+LDH+totalcholesterol+serumamylase+SB+CO2+bloodglucose+urineprotein+CAR,data=hrp,family = binomial())
summary(hrplogisticfit)
hrpY
str(hrpY)
hrpX1Y <- cbind(hrpX1,hrpY)
head(hrpX1Y)
str(hrpX1Y)
hrp2 <- hrp[,c("Ddimer","Ca","LDH","totalcholesterol","serumamylase","SB","CO2","bloodglucose","urineprotein","CAR","type")]
str(hrp2)
hrp2[is.na(hrp2$Ddimer),hrp2$Ddimer]=median(hrp2$Ddimer,na.rm = TRUE)
hrp2$Ddimer[is.na(hrp2$Ddimer)] <- median(hrp2$Ddimer,na.rm = TRUE)
hrp2$Ca[is.na(hrp2$Ca)] <- median(hrp2$Ca,na.rm = TRUE)
hrp2$LDH[is.na(hrp2$LDH)] <- median(hrp2$LDH,na.rm = TRUE)
hrp2$totalcholesterol[is.na(hrp2$totalcholesterol)] <- median(hrp2$totalcholesterol,na.rm = TRUE)
hrp2$serumamylase[is.na(hrp2$serumamylase)] <- median(hrp2$serumamylase,na.rm = TRUE)
hrp2$SB[is.na(hrp2$SB)] <- median(hrp2$SB,na.rm = TRUE)
hrp2$CO2[is.na(hrp2$CO2)] <- median(hrp2$CO2,na.rm = TRUE)
hrp2$bloodglucose[is.na(hrp2$bloodglucose)] <- median(hrp2$bloodglucose,na.rm = TRUE)
hrp2$urineprotein[is.na(hrp2$bloodglucose)] <- median(hrp2$urineprotein,na.rm = TRUE)
hrp2$CAR[is.na(hrp2$CAR)] <- median(hrp2$CAR,na.rm = TRUE)
fix(hrp2)
summary(hrp2$urineprotein)
hrp2$urineprotein[is.na(hrp2$bloodglucose)] <- 0
hrplogisticfit <- glm(type~Ddimer+Ca+LDH+totalcholesterol+serumamylase+SB+CO2+bloodglucose+urineprotein+CAR,data=hrp2,family = binomial())
summary(hrplogisticfit)
library(rms)
library(foreign)
hrplogisticfit2 <- lrm(type~Ddimer+Ca+LDH+totalcholesterol+serumamylase+SB+CO2+bloodglucose+urineprotein+CAR,data=hrp2,x= TRUE,y= TRUE)
summary(hrplogisticfit2)
hrplogisticfit2
install.packages("ROCR")
hrp2$predvalue <- predict(hrplogisticfit2)
library(ROCR)
pred <- prediction(hrp2$predvalue,hrp2$type)
summary(hrp2$urineprotein)
fix(hrp2)
summary(hrp2$urineprotein)
hrp2$urineprotein[is.na(hrp2$urineprotein)] <- 0
summary(hrp2$urineprotein)
hrplogisticfit3 <- lrm(type~Ddimer+Ca+LDH+totalcholesterol+serumamylase+SB+CO2+bloodglucose+urineprotein+CAR,data=hrp2,x= TRUE,y= TRUE)
hrplogisticfit3
hrp2$predvalue <- predict(hrplogisticfit3)
pred <- prediction(hrp2$predvalue,hrp2$type)
perf <- performance(pred,"tpr","fpr")
plot(perf)
auc <- performance(pred,"auc")
auc
hrpauc <- performance(pred,"auc")
hrpauc
hrpauc{y.values}
summary(hrpX1)
summary(hrpx)
q()
hrplogisticfit4 <- lrm(type~Ddimer+Ca+totalcholesterol+serumamylase,data=hrp2,x= TRUE,y= TRUE)
library(rms)
hrplogisticfit4 <- lrm(type~Ddimer+Ca+totalcholesterol+serumamylase,data=hrp2,x= TRUE,y= TRUE)
hrplogisticfit4
hrplogisticfit5 <- lrm(type~Ddimer+Ca+totalcholesterol,data=hrp2,x= TRUE,y= TRUE)
hrplogisticfit5
library(glmnet)
set.seed(123)
hrp.cv.lasso <- cv.glmnet(x=hrpX1,y=hrpY,alpha=1,family="binomal",nfolds = 10)
hrp.cv.lasso <- cv.glmnet(x=hrpX1,y=hrpY,alpha=1,family="binomial",nfolds = 10)
hrp.cv.lasso$lambda.1se
coef(hrp.cv.lasso,s="lambda.1se")
hrplogisticfit6 <- lrm(type~Ddimer+Ca+totalcholesterol+SB+CO2+CAR,data=hrp2,x= TRUE,y= TRUE)
hrplogisticfit6
install.packages("caret")
install.packages("MASS")
library(MASS)
library(rms)
hrplogautofit <- lrm(type~Ddimer+Ca+totalcholesterol+SB+CO2+CAR,data = hrp2,x=TRUE,y=TRUE)
stepAIC(hrplogautofit,direction = "both")
hrplogautofit <- glm(type~Ddimer+Ca+totalcholesterol+SB+CO2+CAR,data = hrp2,family = binomial)
stepAIC(hrplogautofit,direction = "both")
hrplogisticfit7 <- lrm(type~Ddimer+Ca+totalcholesterol+SB,data=hrp2,x= TRUE,y= TRUE)
hrplogisticfit7
anova(hrplogisticfit5,hrplogisticfit6,hrplogisticfit7,test = "Chisq")
anova(hrplogisticfit5,hrplogisticfit6,test = "Chisq")
anova(hrplogisticfit5,hrplogisticfit6,test = "Chisq")
?cv.glmnet
??cv.glmnet
summary(hrp$age)
sd(hrp$age)
summary(hrp)
plot(hrp.cv.lasso)
plot(hrp.cv.lasso,xvar = "lambda")
print(hrp.cv.lasso)
plot(hrp.cv.lasso,xvar = "lambda")
plot(hrp.cv.lasso$glmnet.fit,xvar = "lambda")
library(glmnet)
plot(hrp.cv.lasso$glmnet.fit,xvar = "lambda")
plot(hrp.cv.lasso)
hrplogisticfit5
library(rms)
hrplogisticfit5
hrpcalibrate <- calibrate(hrplogisticfit5,method = "boot",B=1000)
plot(hrpcalibrate)
install.packages("ResourceSelection")
library(ResourceSelection)
hrplogfit5hltest <- hoslem.test(hrplogisticfit5$y,fitted(hrplogisticfit5),g=10)
hrplogisticfit5$y
class(hrplogisticfit5$y)
hrplogfit5yint <- as.integer(hrplogisticfit5$y)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,fitted(hrplogisticfit5),g=10)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,as.integer(fitted(hrplogisticfit5)),g=10)
hrplogfit5yhatint <- as.integer(fitted(hrplogisticfit5))
hrplogfit5yhatint
hrplogfit5yhatnum <- as.numeric(fitted(hrplogisticfit5))
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,hrplogfit5yhatnum,g=10)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,hrplogfit5yhatnum)
hrplogfit5yhatnum
fitted(hrplogisticfit5)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,as.numeric(fitted(hrplogisticfit5)))
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,as.numeric(fitted(hrplogisticfit5)),g=4)
wilcox.test(hrp$PLR~hrp$type)
plot(hrp.cv.lasso)
library(glmnet)
plot(hrp.cv.lasso)
View(hrpX1)
plot(hrp.cv.lasso$glmnet.fit,xvar="lambda")
coef(hrp.cv.lasso,s="lambda.1se")
install.packages("epiDisplay")
plot(perf)
hrpauc
hrp2$predvalue <- predict(hrplogisticfit5)
pred <- prediction(hrp2$predvalue,hrp2$type)
perf <- performance(pred,"tpr","fpr")
plot(perf)
abline(0,1)
mytable <- xtabs(~sex+type,data = hrp)
mytable
chisq.test(mytable)
mytable <- xtabs(~diabetes+type,data = hrp)
mytable
chisq.test(mytable)
mytable <- xtabs(~replase+type,data = hrp)
mytable
chisq.test(mytable)
aggregate(hrp["hct","PTA"]by=list(type=hrp$type),mean)
aggregate(hrp["hct","PTA"],by=list(type=hrp$type),mean)
aggregate(na.omit(hrp["hct","PTA"]),by=list(type=hrp$type),mean)
?mean
?aggregate
aggregate(hrp["hct","PTA"],by=list(type=hrp$type),mean,na.omit=TRUE)
by(hrp["hct","PTA"],hrp$type,mean,na.omit=TRUE)
install.packages("doBy")
library(doBy)
summaryBy(hct+PTA~type,data=hrp,FUN=mean())
summaryBy(hct+PTA~type,data=hrp,FUN=mean)
summaryBy(hct+PTA~type,data=hrp,FUN=mean,na.omit=TRUE)
sumfun <- function(x){c(mean=mean(x, na.rm=TRUE), sd=sd(x, na.rm=TRUE))}
summaryBy(hct+PTA~type,data=hrp,FUN=sumfun)
t.test(hrp$hct~hrp$type)
t.test(hrp$PTA~hrp$type)
wilcox.test(hrp$RBC~hrp$type)
library(doBy)
summaryBy(RBC~type,data=hrp,FUN=fivenum)
summaryBy(RDW~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$RDW~hrp$type)
summaryBy(WBC~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$WBC~hrp$type)
summaryBy(neutrophilicgranulocyte percentage~type,data=hrp,FUN=fivenum)
summaryBy(hrp$neutrophilicgranulocyte.percentage~type,data=hrp,FUN=fivenum)
summaryBy(neutrophilicgranulocyte.percentage~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$neutrophilicgranulocyte.percentage~hrp$type)
summaryBy(Lymphocytepercentage~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Lymphocytepercentage~hrp$type)
summaryBy(Percentage.of.eosinophils~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Percentage.of.eosinophils~hrp$type)
summaryBy(Lymphocyte~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Lymphocyte~hrp$type)
summaryBy(eosinophils~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$eosinophils~hrp$type)
summaryBy(PCT~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$PCT~hrp$type)
summaryBy(CRP~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$CRP~hrp$type)
summaryBy(PCT2~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$PCT2~hrp$type)
library(doBy)
summaryBy(PT~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$PT~hrp$type)
summaryBy(INR~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$INR~hrp$type)
summaryBy(TT~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$TT~hrp$type)
summaryBy(Ddimer~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Ddimer~hrp$type)
summaryBy(Na~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Na~hrp$type)
summaryBy(Ca~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$Ca~hrp$type)
summaryBy(PO4~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$PO4~hrp$type)
summaryBy(HCO3~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$HCO3~hrp$type)
summaryBy(hrp$BUN~type,data=hrp,FUN=fivenum)
summaryBy(BUN~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$BUN~hrp$type)
summaryBy(UricAcid~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$UricAcid~hrp$type)
summaryBy(cystatinC~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$cystatinC~hrp$type)
wilcox.test(hrp$Scr~hrp$type)
library(doBy)
summaryBy(AST~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$AST~hrp$type)
summaryBy(LDH~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$LDH~hrp$type)
summaryBy(hrp$GGT~type,data=hrp,FUN=fivenum)
summaryBy(GGT~type,data=hrp,FUN=fivenum)
wilcox.test(hrp$GGT~hrp$type)
wilcox.test(hrp$cholinesterase~hrp$type)
wilcox.test(hrp$totalcholesterol~hrp$type)
wilcox.test(hrp$TG~hrp$type)
wilcox.test(hrp$LDL~hrp$type)
wilcox.test(hrp$ApolipoproteinB~hrp$type)
wilcox.test(hrp$ApolipoproteinA1Brate~hrp$type)
wilcox.test(hrp$CKMB~hrp$type)
wilcox.test(hrp$serumamylase~hrp$type)
wilcox.test(hrp$PH~hrp$type)
wilcox.test(hrp$PCO2~hrp$type)
wilcox.test(hrp$AB~hrp$type)
wilcox.test(hrp$SB~hrp$type)
wilcox.test(hrp$CO2~hrp$type)
wilcox.test(hrp$SBE~hrp$type)
wilcox.test(hrp$ABE~hrp$type)
wilcox.test(hrp$lacticacid~hrp$type)
wilcox.test(hrp$ionizedcalcium~hrp$type)
wilcox.test(hrp$bloodglucose~hrp$type)
mytable <- xtabs(~urineprotein+type,data = hrp)
mytable
chisq.test(mytable)
fisher.test(mytable)
wilcox.test(hrp$NLR~hrp$type)
wilcox.test(hrp$PLR~hrp$type)
wilcox.test(hrp$MLR~hrp$type)
wilcox.test(hrp$CLR~hrp$type)
wilcox.test(hrp$CAR~hrp$type)
summaryBy(totalcholesterol~type,data=hrp,FUN=fivenum)
summaryBy(TG~type,data=hrp,FUN=fivenum)
summaryBy(LDL~type,data=hrp,FUN=fivenum)
summaryBy(ApolipoproteinB~type,data=hrp,FUN=fivenum)
summaryBy(ApolipoproteinA1Brate~type,data=hrp,FUN=fivenum)
summaryBy(CKMB~type,data=hrp,FUN=fivenum)
summaryBy(serumamylase~type,data=hrp,FUN=fivenum)
summaryBy(PH~type,data=hrp,FUN=fivenum)
summaryBy(PCO2~type,data=hrp,FUN=fivenum)
summaryBy(AB~type,data=hrp,FUN=fivenum)
summaryBy(SB~type,data=hrp,FUN=fivenum)
summaryBy(hrp$CO2~type,data=hrp,FUN=fivenum)
summaryBy(CO2~type,data=hrp,FUN=fivenum)
summaryBy(SBE~type,data=hrp,FUN=fivenum)
summaryBy(ABE~type,data=hrp,FUN=fivenum)
summaryBy(lacticacid~type,data=hrp,FUN=fivenum)
summaryBy(ionizedcalcium~type,data=hrp,FUN=fivenum)
summaryBy(bloodglucose~type,data=hrp,FUN=fivenum)
summaryBy(NLR~type,data=hrp,FUN=fivenum)
summaryBy(PLR~type,data=hrp,FUN=fivenum)
summaryBy(MLR~type,data=hrp,FUN=fivenum)
summaryBy(CLR~type,data=hrp,FUN=fivenum)
summaryBy(CAR~type,data=hrp,FUN=fivenum)
library(rms)
hrpfitcoef <- coef(hrplogisticfit5)
se <- c(1.5088,0.0003,0.6803,0.0326)
hrpfit5coefci <- cbind(hrpfitcoef-1.96*se,hrpfitcoef+1.96*se)
hrpfit5OR <-exp(cbind("OR"=hrpfitcoef,"LL"=hrpfit5coefci[,1],"UL"=hrpfit5coefci[,2])
hrpfit5OR <-exp(cbind("OR"=hrpfitcoef,"LL"=hrpfit5coefci[,1],"UL"=hrpfit5coefci[,2])
hrpfit5OR <-exp(cbind("OR"=hrpfitcoef,"LL"=hrpfit5coefci[,1],"UL"=hrpfit5coefci[,2]))
hrpfit5OR
library(dplyr)
library(pROC)
hrproc <- roc(predvalue~type,data=hrp2)
hrp2$predvalue
hrproc <- roc(hrp2$type,hrp2$predvalue)
auc(hrproc)
ci(auc(hrproc))
library(ResourceSelection)
hrplogfit5hltest <- hoslem.test(hrplogisticfit5$y,fitted(hrplogisticfit5),g=10)
hrplogisticfit5$y
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,fitted(hrplogisticfit5),g=10)
hrplogfit5yint <- as.integer(hrplogisticfit5$y)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,fitted(hrplogisticfit5),g=10)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,hrplogfit5yhatnum,g=10)
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,as.numeric(fitted(hrplogisticfit5)))
hrplogfit5hltest <- hoslem.test(hrplogfit5yint,as.numeric(fitted(hrplogisticfit5)),g=4)
hoslem.test(hrplogisticfit5$y,fitted(hrplogisticfit5))
hoslem.test(hrplogisticfit5$y,as.numeric(fitted(hrplogisticfit5)))
hoslem.test(hrplogisticfit$y,fitted(hrplogisticfit),g=10)
hrpglmfit5 <- glm(type~Ddimer+Ca+totalcholesterol,data=hrp2,family = binomial())
library(pROC)
hrp2$glmpredvalue <- predict(hrpglmfit5)
hrpglmroc <- roc(hrp2$type,hrp2$glmpredvalue)
auc(hrpglmroc)
hrpfit5OR
hoslem.test(hrpglmfit5$y,fitted(hrpglmfit5),g=10)