Representative sample as a LP problem
Representative sample as a LP problem

Summary/Abstract: Empirical analyses usually estimate the effect of some variables, considered as independent, on one (or more), considered as dependend. It is rare situation to have the whole population available for the analysis. Then, usually a subsample is chosen and the conclusions are generalized to the whole population. This approach is based on the concept that the sample represents, in some meaning the population. If a sample is large enough and the selection process is random, usually the sample is considered to be representative one. However in some cases the sample should satisfy some restrictions on selected characteristics, which are previously known for the population. Here we propose an approach to this problem as a solution of a task of integer linear programming.

• Details
• Contents