Multiple regression analysis method for mixed data

Aiming at the mixed data composed of complete data, type-I censored data and type-II censored data which are very common in accelerated life test, a multiple regression analysis method were developed. The multiple regression model was established, and the best unbiased estimations of regression coefficients and the standard deviation were given. Furthermore, the point estimation and confidence limits of percentiles were obtained. The presented method is applicable to not only the normal distribution but also the Weibull distribution, the extreme value distribution and some other location-scale distributions. The traditional multiple regression analysis method which is strictly suitable to the complete data is extended to the mixed data composed of complete data and censored data. Compared with the traditional method, the presented method can reach a high precision and a great number of specimens can be saved.