Estimating Missing Data with Bayes Bootstrap Regression Imputation
Abstract
This research is about estimating missing data when dependent variable Y is correlated with independent variable X, and X and Y are distributed as normal. The proposed method for estimating missing data is Bayes-Bootstrap regression imputation method (BRI) that is compared with regression imputation method (RI) and distance regression imputation method (DRI). The measurement criteria is mean absolute error (MAE). Comparing of estimating missing data used the Monte Carlo simulation technique. The results of study indicate that BRI and RI are more accuracy than DRI for all cases, but BRI presents the lowest mean absolute error in some case. Therefore, researchers introduce the BRI method for estimating missing data when the correlation coefficient between dependent variable Y and independent variable X is known and both variable distributions are normal distributions. Keywords : : missing data ; Bayes-Bootstrap regression imputation ; Distance regression imputation ; regression imputationReferences
Brick, J.M., and Kalton, G. (1996). Handling missing data in survey research. Statistical Methods in Medical
Research, 5(3), 215-238.
Chaimongkol, W. (2005). Three composite imputation methods for item nonresponse estimation in sample
surveys(Doctoral dissertation) Graduate School of Applied Statistics, National Institute of Development
Administration, Bangkok.
Jitthavech, J. (2015). Regression Analysis (1st ed.). Bangkok, Thailand: Academic Promotion and Development
Program, National Institute of Development Administration. (in Thai)
Lin, J. Q., Wu, H. C., Chan, S. C. (2017). A new regularized recursive dynamic factor analysis with variable
forgetting factor for wireless sensor networks with missing data. IEEE International Symposium on
Circuits and Systems, 1–4.
Merlise, A. C., Herbert K. H. L. (2000). Bagging and the Bayesian Bootstrap. Retrieved Jan, 2019, form
https://www.researchgate.net/publication/2469163.
Peng, C.Y.J., Harwell, M., Liou, S.M., Ehman. LH. (2006) Advances in missing data methods and implications for
educational research. Real data analysis, 31–78.
Pimchanok, C., Watchareeporn, C. (2017). A comparison of the estimation methods for missing data in sample
survey. The Journal of Applied Science, 16(1), 60-73.
Rubin, D. (1981). The Bayesian bootstrap, Annals of Statistics, 9, 130-134.
Troyanskaya, O., Cantor, M., Sherlock, G. (2001). Missing value estimation methods for DNA
microarrays. Bioinformatics, 17(6), 520–525, 2001.
Research, 5(3), 215-238.
Chaimongkol, W. (2005). Three composite imputation methods for item nonresponse estimation in sample
surveys(Doctoral dissertation) Graduate School of Applied Statistics, National Institute of Development
Administration, Bangkok.
Jitthavech, J. (2015). Regression Analysis (1st ed.). Bangkok, Thailand: Academic Promotion and Development
Program, National Institute of Development Administration. (in Thai)
Lin, J. Q., Wu, H. C., Chan, S. C. (2017). A new regularized recursive dynamic factor analysis with variable
forgetting factor for wireless sensor networks with missing data. IEEE International Symposium on
Circuits and Systems, 1–4.
Merlise, A. C., Herbert K. H. L. (2000). Bagging and the Bayesian Bootstrap. Retrieved Jan, 2019, form
https://www.researchgate.net/publication/2469163.
Peng, C.Y.J., Harwell, M., Liou, S.M., Ehman. LH. (2006) Advances in missing data methods and implications for
educational research. Real data analysis, 31–78.
Pimchanok, C., Watchareeporn, C. (2017). A comparison of the estimation methods for missing data in sample
survey. The Journal of Applied Science, 16(1), 60-73.
Rubin, D. (1981). The Bayesian bootstrap, Annals of Statistics, 9, 130-134.
Troyanskaya, O., Cantor, M., Sherlock, G. (2001). Missing value estimation methods for DNA
microarrays. Bioinformatics, 17(6), 520–525, 2001.
Downloads
Published
2021-05-05
Issue
Section
Research Article
