What methods would you recommend for testing whether a variable Y is the weighted sum of n, not necessarily independent, variables? Only m < n of these variables are known.
The problem is easy if the variables are independent. (In my setting their co-variance is close to zero).