Simulated Data Generation Using DAG Causal Structure Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for generating simulated data fail to accurately mimic real data in terms of causal and correlation relationships.
Innovation Solution
A method involving a directed acyclic graph (DAG) is used to learn the topological order of variables, compute parameter estimates, and generate simulated data that preserves causal and correlation relationships with real data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current techniques are used to generate simulated data, then the generation process is simple, but the simulated data fails to accurately preserve causal and correlation relationships with real data
Solution Approach 1:
The patent segments the data generation process into distinct components: learning the topological order of variables from real data, computing parameter estimates, generating error values, and synthesizing simulated data. This segmentation allows each component to be optimized independently, improving the preservation of causal relationships while managing overall complexity.
Solution Approach 2:
The patent performs preliminary actions by first learning the topological order and computing parameter estimates from real data before generating simulated data. This preliminary analysis of causal structures enables the simulated data to accurately preserve relationships without requiring complex real-time processing during data synthesis.
2Measurement precision
If complex methods are used to preserve causal relationships, then the accuracy of simulated data improves, but the computational complexity increases
Solution Approach 1:
The patent creates a simplified copy of the causal structure by learning the topological order from real data and representing it as a directed acyclic graph. This copied structure can be efficiently sampled to generate simulated data, avoiding the need to repeatedly analyze complex real-world relationships during each data generation process.
Solution Approach 2:
The patent transforms the complex problem of preserving causal relationships into parameter estimation and sampling from a learned topological order. By changing the representation from raw data relationships to structured parameters (topological order, parameter estimates, error values), the system achieves high accuracy with reduced computational overhead during data generation.
Data Source
AI summary
A system and method include learning a topological order of a plurality of variables in a directed acyclic graph based on real data, computing parameter estimate values corresponding to the real data, computing error values based on the real data, the topological order, and the parameter estimate values, generating simulated data from the parameter estimate values and the error values, such that simulated variables in the simulated data preserve a causal relationship between variables in the real data, and the simulated variables in the simulated data preserve a correlation relationship between the variables in the real data, and reorganizing and outputting the simulated data based on the topological order.