synthetic data for logistic regression

synthetic_data.R

# --- synthetic data for logistic regression # 
# Sean Chang 2014 
# sean.chang@duke.edu

# sample size 
n = 1e4  
x <- rnorm(n, -1, 0.5) + rnorm(n,0.5,0.5) 

# transfer x to [0,1]
x = exp(x)/(1+exp(x))
hist(x)
mean(x)

# generate positive coefficients
beta = 1 + rnorm(n, 0, 0.375) #noise

# generate y
prob = exp(x*beta)/(1+exp(x*beta))
plot(x, prob)
y = rbinom(n, 1, prob)

# save file
#save(x, y, file="/v_share/simulated.RData")
df = data.frame(x,y)
write.table(df,file="/v_share/simulated.csv", sep = ",")