library(RMySQL)
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="127.0.0.1", port=3307, password='kuancfit')
data <- dbGetQuery(mycon, "select * from corr_drug_drug")
compute.p.values <- function(row) {
    x <- c(rep(1,row[4]), rep(1,row[5]), rep(0,row[6]), rep(0,row[7]))
    y <- c(rep(1,row[4]), rep(0,row[5]), rep(1,row[6]), rep(0,row[7]))
    return(cor.test(x,y)$p.value)
}
data$pvalues <- apply(data, 1, compute.p.values)
library(RMySQL)
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="localhost", port=3306, password='kuancfit')
data <- dbGetQuery(mycon, "select * from corr_drug_drug")
compute.p.values <- function(row) {
    x <- c(rep(1,row[4]), rep(1,row[5]), rep(0,row[6]), rep(0,row[7]))
    y <- c(rep(1,row[4]), rep(0,row[5]), rep(1,row[6]), rep(0,row[7]))
    return(cor.test(x,y)$p.value)
}
data$pvalues <- apply(data, 1, compute.p.values)
head(data)
data <- dbGetQuery(mycon, "select * from corr_drug_drug_old")
data$pvalue <- apply(data, 1, compute.p.values)
head(data)
?write.csv
write.csv(data, "~/corr_drug_drug_011012.csv", row.names=F, quote=F)
data <- dbGetQuery(mycon, "
select o.z_score1 as offsides, s.z_score1 as sider, shared_indication is not null as shared
from drug_similarity_b o
join drug_similarity_sider s using (stitch_id1, stitch_id2)
left join drug_shared_indication using (stitch_id1, stitch_id2)
")
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="localhost", port=3306, password='kuancfit')
data <- dbGetQuery(mycon, "
select o.z_score1 as offsides, s.z_score1 as sider, shared_indication is not null as shared
from drug_similarity_b o
join drug_similarity_sider s using (stitch_id1, stitch_id2)
left join drug_shared_indication using (stitch_id1, stitch_id2)
")
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="localhost", port=3306, password='kuancfit')
data <- dbGetQuery(mycon, "
select o.z_score1 as offsides, s.z_score1 as sider, shared_indication is not null as shared
from drug_similarity_b o
join drug_similarity_sider s using (stitch_id1, stitch_id2)
left join drug_shared_indication using (stitch_id1, stitch_id2)
")
head(data)
sum(is.na(data))
sum(is.na(data$offsides))
data <- data[!is.na(data$offsides),]
head(data)
fit <- glm(shared ~ offsides + sider, data, family="binomial")
summary(fit)
pred <- prediction(predict(fit), data$shared)
performance(pred, measure="auc")@y.values[[1]]
library(ROCR)
pred <- prediction(predict(fit), data$shared)
performance(pred, measure="auc")@y.values[[1]]
install.packages('ROCR')
library(ROCR)
pred <- prediction(predict(fit), data$shared)
performance(pred, measure="auc")@y.values[[1]]
perf <- performance(pred, measure="tpr", x.measure="fpr")
perf
names(perf)
head(perf)
?perf
?performance
head(perf)
head(perf$x)
write.csv(perf, '~/Dropbox/performance.csv')
head(perf$x.values)
head(perf#x.values)
#
)
head(perf@x.values)
write.csv(cbind(perf@x.values, perf@y.values), '~/Dropbox/performance.csv', row.names=F, quote=F)
x <- cbind(perf@x.values, perf@y.values)
head(x)
head(as.numeric(perf@x.values))
head(as.data.frame(perf@x.values))
head(as.matrix(perf@x.values))
head(as.matrix(perf@y.values))
head(as.matrix(perf@y.values[[1]]))
head(perf@y.values[[1]])
write.csv(cbind(perf@x.values[[1]], perf@y.values[[1]]), '~/Dropbox/performance.csv', row.names=F, quote=F)
write.csv(cbind(perf@x.values[[1]], perf@y.values[[1]]), '~/Dropbox/performance.csv', row.names=F, col.names=F, quote=F)
pred <- prediction(data$sider, data$shared)
performance(pred, measure="auc")@y.values[[1]]
pred <- prediction(data$offsides, data$shared)
performance(pred, measure="auc")@y.values[[1]]
pred <- prediction(predict(fit), data$shared)
performance(pred, measure="auc")@y.values[[1]]
summary(fit)
pred <- prediction(data$sider, data$shared)
perf <- performance(pred, measure="tpr", x.measure="fpr")
write.csv(cbind(perf@x.values[[1]], perf@y.values[[1]]), '~/Dropbox/performance.csv', row.names=F, quote=F)
pred <- prediction(data$offsides, data$shared)
perf <- performance(pred, measure="tpr", x.measure="fpr")
write.csv(cbind(perf@x.values[[1]], perf@y.values[[1]]), '~/Dropbox/performance.csv', row.names=F, quote=F)
summary(fit)
pred <- prediction(data$sider, data$shared)
performance(pred, measure="auc")@y.values[[1]]
pred <- prediction(data$offsides, data$shared)
performance(pred, measure="auc")@y.values[[1]]
pred <- prediction(predict(fit), data$shared)
performance(pred, measure="auc")@y.values[[1]]
summary(fit)
head(data)
data <- dbGetQuery(mycon, "
select stitch_id1, stitch_id2, o.z_score1 as offsides, s.z_score1 as sider, shared_indication is not null as shared
from drug_similarity_b o
join drug_similarity_sider s using (stitch_id1, stitch_id2)
left join drug_shared_indication using (stitch_id1, stitch_id2)
where o.z_score1 is not null
and s.z_score1 is not null
")
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="localhost", port=3306, password='kuancfit')
library(RMySQL)
mycon <- dbConnect(MySQL(), user='root', dbname="project_aers", host="localhost", port=3306, password='kuancfit')
data <- dbGetQuery(mycon, "
select stitch_id1, stitch_id2, o.z_score1 as offsides, s.z_score1 as sider, shared_indication is not null as shared
from drug_similarity_b o
join drug_similarity_sider s using (stitch_id1, stitch_id2)
left join drug_shared_indication using (stitch_id1, stitch_id2)
where o.z_score1 is not null
and s.z_score1 is not null
")
fit <- glm(shared ~ offsides + sider, data, family="binomial")
summary(fit)
data$both <- predict(fit)
head(data)
write.csv(data, '~/drug_similarity_bsider_logreg_indication.csv', row.names=F, header=F, quote=F)
write.csv(data, '~/drug_similarity_bsider_logreg_indication.csv', row.names=F, quote=F)