Orthogonal linear regression algorithm based on augmented matrix formulation

Scope and Purpose : In this paper, a new technique for solving a multivariate linear model using the orthogonal least absolute values regression is proposed. The orthogonal least absolute values (ORLAV) regression minimises the sum of the absolute, orthogonal distance from each data point to the resulting regression hyperplane. In a large set of equations where the variables are independent of each other, it is most convenient to measure the orthogonal distances. The other advantage of ORLAV regression is that this method is insensitive to outliers if minimising the squared distances, for example, an extreme outlier can be most significant.