Columnwise neural imputation for incomplete ordinal psychometric data.

Abstract

COLNI (pronounced “col-knee”) is an ANN-based framework for imputing missing values in tabular data, with a particular focus on ordinal response scales frequently encountered in education, psychology, behavioral science, and related domains. Extensive simulation studies and real-world evaluations---summarized in our paper, Columnwise Neural Imputation for Incomplete Ordinal Psychometric Data---validate COLNI’s effectiveness across diverse conditions