ROBUST MODELING WITH ERRATIC DATAt

An attractive alternative to least-squares data determined by using the median rather than the modeling techniques is the use of absolute value arithmetic mean. Algorithms for absolute error error criteria. Unlike the least-squares techniques minimization are often approximately as costly the inclusion of some infinite blunders along with as least-squares algorithms; however, unlike the data will hardly affect the solution to an least-squares, they naturally lend themselves to otherwise well-posed problem. An example of inequality or bounding constraints on models. this great stability is seen when an average is