Sensitivity of ridge-type estimation methods to condition number

摘要

An analysis of ridge-type estimation methods and their performance in response to different levels of ill-conditioning is presented. By adjusting the observation accuracy and the sampling rate, the amount of ill-conditioning in a linear estimation problem is altered. The effectiveness of the ridge-type methods in response to these changes is evaluated by comparing the solution accuracy with those of the standard minimum variance techniques. Results show the improved performance of the ridge-type estimation method for various levels of ill-conditioning. In addition, the amount of improvement increases as the level of ill-conditioning present in the problem is increased.