Activity Number:
|
419
- Contributed Poster Presentations: Government Statistics Section
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Government Statistics Section
|
Abstract #302980
|
|
Title:
|
Reconstructing Matrices with Linear Programming
|
Author(s):
|
Luis Frank*
|
Companies:
|
Universidad de Buenos Aires
|
Keywords:
|
matrix adjustment;
matrix balancing;
matrix reconciliation;
linear programming;
official statistics;
National Accounts
|
Abstract:
|
The article explains how to adjust, balance and temporarily disaggregate matrices (to "reconstruct" matrices in a broad sense) with linear programming. The main advantage of this method compared to RAS (standard method recommended in National Accounts manuals) is that it allows reconstructing matrices even with missing data in the benchmark matrix, or with marginal vectors that are inconsistent with each other. Reconstructing matrices with linear programming also allows enriching the solution by incorporating alternative or complementary information to the adjustment process. The article also includes numerical examples with small matrices to show the performance of the proposed method.
|
Authors who are presenting talks have a * after their name.