用“离差绝对值之和最小”求回归直线方程

To Evaluate the Regression line Equation by the Way of the Least sum of Deviation Absolate Value

本文从“离差绝对值之和最小”回归方程会有特殊运用价值的观点出发,介绍了用作图,操作、计算相结合的方法,求任何一批样点的“离差绝对值之和最小”回归方程.同时对直线转动中,样点离差绝对值变化的规律进行了讨论.取得了直线转动与总体样点离差绝对值变化的关系式和有关性质,证明了“离差绝对值之和最小”回归方程求算方法的可行性和科学性.

The essay begins with the view of .special usage of the regression equation that the sum of deviation absolate value is the least.by means of drawing.operating and calculating to evaluate the random sampled point regression equation that its sum of the deviation absolate value is the least.meanwhile the essay analyses the rule that the sampled point deviation absolate value changes,while the, line rotating.which acquires the form and relevant character between the line rotating and the total sampled point deviation absolate value changing.and proves the feasibility and science of the evaluation of the regression equation that its sum of the deviation absolate value is the least.