DML-PM4

[alexanderquispe]

• https://github.com/DoubleML/doubleml-for-py
• https://github.com/microsoft/EconML
• https://humboldt-wi.github.io/blog/research/applied_predictive_modeling_19/double_machine_learning/

[alexanderquispe]

• How can we get LASSO post-selection method using sklearn? @anzonyqr

[anzonyquispe]

# Treatment Variable
D     <- rdata[which(colnames(rdata) == d)]

# Outcome Variable
Y     <- rdata[which(colnames(rdata) == y)]

# Construct matrix Z

Z <- rdata[which(colnames(rdata) %in% c(X1,X2,census))]

dim(Z)

clu <- rdata[which(colnames(rdata) == "CountyCode")] #for clustering the standard errors
data <- data.frame(cbind(Y, D, Z,as.matrix(clu)))

I believe all these lines of code are not necessary since they are just changing the order of the columns, but they do not make any change in their values. Additionally, we have to consider that our output rdata from cleaning data is equal to data r-dataframe at 6 decimals.