Timezone: »
MISNN: Multiple Imputation via Semi-parametric Neural Networks
Qi Long · Yiliang Zhang · Zongyu Dai · Zhiqi Bu
Sat Jul 24 03:10 PM -- 03:15 PM (PDT) @
Multiple imputation (MI) has been widely applied to missing value problems in biomedical, social and econometric research, in order to avoid improper inference in the downstream data analysis. In the presence of high-dimensional data, imputation models that include feature selection, especially $\ell_1$ regularized regression (such as Lasso, adaptive Lasso and Elastic Net), are common choices to prevent the model from underdetermination. However, conducting MI with feature selection is difficult: existing methods are often computationally inefficient and poor in performance. We propose MISNN, a novel and efficient algorithm that incorporates feature selection for MI. Leveraging the approximation power of neural networks, MISNN is a general and flexible framework, compatible to any feature selection method, any neural network architecture, high/low-dimensional data and general missing patterns. Through empirical experiments, MISNN has demonstrated great advantages over state-of-the-art imputation methods (e.g. Bayesian Lasso and matrix completion), in terms of imputation accuracy, statistical consistency and computation speed.
Author Information
Qi Long (University of Pennsylvania)
Yiliang Zhang (University of Pennsylvania)
Zongyu Dai (University of Pennsylvania)
Zhiqi Bu (University of Pennsylvania)
More from the Same Authors
-
2021 : MISNN: Multiple Imputation via Semi-parametric Neural Networks »
Zhiqi Bu · Zongyu Dai · Yiliang Zhang · Qi Long -
2021 : Differentially Private Bayesian Neural Network »
Zhiqi Bu · Qiyiwen Zhang · Kan Chen · Qi Long -
2021 : On the Convergence of Deep Learning with Differential Privacy »
Zhiqi Bu · Hua Wang · Qi Long · Weijie Su -
2021 : Fairness in Missing Data Imputation »
Yiliang Zhang · Qi Long -
2023 Poster: Differentially Private Optimization on Large Model at Small Cost »
Zhiqi Bu · Yu-Xiang Wang · Sheng Zha · George Karypis -
2021 Poster: Accuracy, Interpretability, and Differential Privacy via Explainable Boosting »
Harsha Nori · Rich Caruana · Zhiqi Bu · Judy Hanwen Shen · Janardhan Kulkarni -
2021 Spotlight: Accuracy, Interpretability, and Differential Privacy via Explainable Boosting »
Harsha Nori · Rich Caruana · Zhiqi Bu · Judy Hanwen Shen · Janardhan Kulkarni