International Journal of Applied Science and Technology

ISSN 2221-0997 (Print), 2221-1004 (Online) 10.30845/ijast

Determination of the Best Imputation Technique for Estimating Missing Values when Fitting the Weibull Distribution
Ahmad Shukri Yahaya, Nor Azam Ramli, Fauziah Ahmad, Norlida Mohd Nor, Muhammad Nor Hakim Bahrim

Abstract
Missing values are common in any scientific work. Nine single imputation techniques were compared to determine the best technique for estimating missing values when fitting Weibull distributions. Simulation technique was used to obtain random variables for the Weibull distributions. Samples of sizes 30, 50, 200 and 300 were used to simulate the Weibull distribution. To determine the best imputation technique, four error measures were used that are the normalized absolute error, root mean square error, index of agreement and root mean square error. This study shows that no single imputation technique is the best for each sample size and for each percentage of missing values.

Full Text: PDF