TY - JOUR
TI - Research and comparison of imputation methods of three parameter Weibull distribution with missing warp breakage data
AU - Xi Meng-Yao
AU - Lai Jun-Feng
JN - Thermal Science
PY - 2026
VL - 30
IS - 2
SP - 941
EP - 950
PT - Article
AB - Warp breakage data is a critical metric for assessing the durability of warp  yarns, which typically follows a Weibull distribution. The three-parameter  Weibull distribution is a prevalent distribution in which scale, position,  and shape parameters influence the statistical information of the  distribution. The confidence interval is the degree to which the true value  has a certain probability of falling around the measurement result, and it  is important statistical information. In instances where the calculation of  the confidence interval of a Weibull distribution is performed in the  presence of missing data, employing solely complete data estimation can  lead to the introduction of bias. However, in the context of missing data,  the utilization of bootstrap interpolation has been demonstrated to yield  outcomes that are often superior to those obtained through unprocessed data.  The specific process entails the simulation of the Weibull dataset,  incorporating random missing values of 5%, 20%, 50%, and 75%, and the  subsequent comparison of the results of calculating confidence intervals  using not missing and Bootstrap interpolation. The experimental results  demonstrate that the presence of missing data exerts a negligible influence  on the extent of confidence intervals, and the Bootstrap interpolation  method yields values that approximate the median. This phenomenon results in  an overall scarcity of discrete data. The experiment was verified using real  warp fracture data, and the results were essentially consistent with the  simulation experiment.
DO - 10.2298/TSCI2602941X
ER -