THE DEVELOPMENT ESTIMATING MISSING DATA METHOD FOR REPEATED MEASUREMENTS BY TWO-WAY IMPUTATION AND FIRST-ORDER AUTOREGRESSIVE MODEL
Article Sidebar
Main Article Content
Abstract
This research aimed to develop a method estimate a missing data for repetition by two-way imputation and First-Order autoregressive model (IWTWAR (1)) and to compare the efficiency of estimation of missing data by comparing the estimation of missing data by using the traditional method. The scope of the study was a number of repeat measurements divided into 3 and 6 times. The size of the sample was divided into 10, 20 and 50 and the number of percent of missing data was divided into 5, 10, 15 and 20 percent. The researcher considered the efficiency of the missing data estimating method by using mean squared error (MSE). The result showed that the estimating of missing data by two-way imputation and the first-order autoregressive model for the most effective if repetition 6 times. Considering the developmental scores in a Linear Growth Model and the developmental scores were equal to 0.5 and 2.5. When the sample sizes were equal 10 and 20 units, missing data were 5, 10, 15 and 20 percent, in case a linear and non-linear score. The researcher developed the efficient estimating methods varying with different methods in the case of developmental point’s equal 0.5 and were significantly different from the missing data estimating method with the First-Order of autoregressive model (AR (1)) and using missing data estimating methods by mean substitution (MI). The case repetitions 3 times estimating missing data with the First-Order of the autoregressive model showed the highest efficiency in all cases.
Article Details
References
Wongprasert, S. (2011). The development of the method of estimating the lost data by loop weighted using the method of jackknife and regression analysis. Doctorate of Philosophy, Maha Sarakham University. (in Thai).
วราฤทธิ์ พานิชกิจโกศลกุล. (2554). การเปรียบเทียบวิธีการพยากรณ์ข้อมูลล่วงหน้าหนึ่งคาบเวลาสำหรับตัวแบบอัตตสหสัมพันธ์อันดับที่หนึ่ง. วารสารวิทยาศาสตร์ ม. อุบลฯ ฉบับพิเศษ, 1(มกราคม-ธันวาคม), 1–9.
Phanitchakitkosolkun, W. (2011). A comparison of the one time period forward forecasting method for first order correlation models. Special edition Ubon Journal of Science Journal, 1 (January-December), 1–9. (in Thai).
วิชิต หล่อจีระชุณหุกุล. (2539). เทคนิคการพยากรณ์. กรุงเทพมหานคร: สถาบันบัณฑิตพัฒนบริหารศาสตร์.
Lorjirunhoonkul, W. (1996). Forecasting techniques. Bangkok: National Institute of Development Administration. (in Thai).
วิชิต หล่อจีระชุณหุกุล และจิราวัลย์ จิตรถเวช. (2548). เทคนิคการพยากรณ์. พิมพ์ครั้งที่ 3. กรุงเทพมหานคร: สถาบันบัณฑิตพัฒนบริหารศาสตร์.
Lorjirunhoonkul, W., & Jitaravej, J. (2005). Forecasting Techniques. 3rd edition, Bangkok: The National Institute of Development Administration. (in Thai).
สุพรรณี อึ้งปัญสัตวงศ์. (2555). เทคนิคการพยากรณ์เชิงสถิติ. ขอนแก่น: คณะวิทยาศาสตร์มหาวิทยาลัยขอนแก่น.
Ungpunsatwong, W. (2012). Statistical forecasting techniques. Khon Kaen: Faculty of Science, Khon Kaen University. (in Thai).
Allison. (2012). Handling Missing Data by Maximum Likelihood. Statistics and Data Analysis, 312(2012), 1145–1152.
Bernaards, C.A., & Sijtsma, K. (2000). Influence of imputation and EM methods on factor analysis when item nonresponse in questionnaire data is nonignorable. Multivariate Behav, 3(5), 321–364.
Cismondi & other. (2013). Weighted Nearest Neighbor and Regression Imputation. In Paper Presented at the Meeting of 9th Asia-Pacific Decision Sciences Institute. Korea: APDSI-KOPOMS Seoul.
DeSilvio, M. (2009). A Variance Ratio Statistics for Assessing the Missing Data Mechanism: An Empirical Study. Dissertation Abstracts International, 63(1), 5698–A.
Garrett, M, F., Nan M.L., & James. H. W. (2011). Applied Longitudinal Analysis. New Jersey: Wiley& sons.
Ginkel, J. R. V. (2007). Two-way imputation: A Bayesian method for estimating missing scores in tests and questionnaires, and an accurate approximation. Computational Statistics & Data Analysis, 51(2007), 4013 – 4027.
Heeringa, S. G. (2000). Multivariate Imputation of Coarsened Survey on Household Wealth. Diddertation Abstracts International, 61(1), 1112–1120.
Little, R. J. and Rubin, D. (2002). Statistical Analysis with Missing Data (2nd ed.). New York: John Wiley and Sons.
Lynette H. & Murray J. (2003). Mixture model clustering for mixed data with missing information. Computational Statistics & Data Analysis, 41(2003), 429 – 440.
Rencher & Christensen. (2012). Method of Multivariate Analysis (3rd ed.). New York: Wiley Series.
Schafer, J. L., & Schenker, N. (2000). Inference with Imputed Conditional Means. Journal of the American Statistical Association, 95(March), 144–154.
So, B. S., & Shin, D. (1999). Recursive Mean Adjustment in Time Series Infer-ences. Statistics and Probability Letters, 4(3), 65–73.
