中途打ち切りデータによる回帰分析の新しい推定方法

概要

論文の詳細を見る
Suppose there are two mechanism A and B which affect the way of variation in a quality characteristic y, and only the minimum value of y_A by mechanism A and y_B by mechanism B can be observed. The problem is to analyze the cause and effect relationship between the quality characteristic y and factors x. For the analysis of these relationship, the ordinary least squares method on data stratified according to the mechanism A or B is conveniently used, or the method which developed for failure time analysis can be applied. The former, however, is poor in the accuracy of estimated parameters, and the latter has difficulties in computation. This paper proposes a new simple method by correcting biases based on the estimation of censoring rate. It also evaluates the accuracy of the proposed method in addition to other methods. The proposed method is simpler and has less deficiency in the accuracy of estimates than other methods.
社団法人日本品質管理学会の論文
1989-04-15