A new type of estimator has been developed to better predict coefficients in a mixed effect linear model when dealing with repeated-measures data. This estimator, called d*ω, is more effective than other existing methods like GLS and Ridge estimators. It helps handle issues like multiconllinearity and improves the accuracy of predictions in these types of models.