Converting general problem to distance-from-origin problem
Suppose we wish to find the nearest point on a plane to the point (), where the plane is given by . We define , , , and , to obtain as the plane expressed in terms of the transformed variables. Now the problem has become one of finding the nearest point on this plane to the origin, and its distance from the origin. The point on the plane in terms of the original coordinates can be found from this point using the above relationships between and , between and , and between and ; the distance in terms of the original coordinates is the same as the distance in terms of the revised coordinates.
The formula for the closest point to the origin may be expressed more succinctly using notation from linear algebra. The expression in the definition of a plane is a dot product, and the expression appearing in the solution is the squared norm. Thus, if is a given vector, the plane may be described as the set of vectors for which and the closest point on this plane is the vector
In either the coordinate or vector formulations, one may verify that the given point lies on the given plane by plugging the point into the equation of the plane.
To see that it is the closest point to the origin on the plane, observe that is a scalar multiple of the vector defining the plane, and is therefore orthogonal to the plane.
Thus, if is any point on the plane other than itself, then the line segments from the origin to and from to form a right triangle, and by the Pythagorean theorem the distance from the origin to is
Since must be a positive number, this distance is greater than , the distance from the origin to .
Alternatively, it is possible to rewrite the equation of the plane using dot products with in place of the original dot product with (because these two vectors are scalar multiples of each other) after which the fact that is the closest point becomes an immediate consequence of the Cauchy–Schwarz inequality.
Closest point and distance for a hyperplane and arbitrary point
The vector equation for a hyperplane in -dimensional Euclidean space through a point with normal vector is or where .
The corresponding Cartesian form is where .
The closest point on this hyperplane to an arbitrary point is