Super-relevant synthetic data with individualized Shapley values

Abstract

Individualized inference (or prediction) is an approach to data analysis that provides tailored analytical insights for specific queries. It is increasingly relevant thanks to the availability of large datasets. This paper presents an algorithm that identifies relevant observations through similarity metrics and further refines this selection by weighting with Shapley values. The probability distribution over thisselection allows for generating synthetic controls, which in turn can be used to generate a robust inference (or prediction). Data collected from repeating this procedure for different queries provides a deeper understanding of the general process that generates the data.