Abstract
Trends and regression lines are almost always computed via the least-squares method. A conceptually simpler approach, resulting in geometrically true “best fit,” is illustrated here.
Get full access to this article
View all access options for this article.