Imputation for Repeated Bounded Outcome Data: Statistical and Machine-Learning Approaches

Real-life data are bounded and heavy-tailed variables. Zero-one-inflated beta (ZOIB) regression is used for modelling them. There no appropriate methods to address the problem of missing in repeated outcomes. We developed an imputation method using ZOIB (i-ZOIB) compared its performance with those naïve machine-learning methods, different distribution shapes settings designed simulation study. The was measured employing absolute error (MAE), root-mean-square-error (RMSE) unscaled mean relative (UMBRAE) methods. results varied depending on missingness rate mechanism. i-ZOIB ANN, SVR RF showed best performance.