Least Squares

Definition and Derivations

We have already spent much time finding solutions to 

        Ax  =  b 

If there isn't a solution, we attempt to seek the x that gets closest to being a solution.  The closest such vector will be the x such that

        Ax  =  proj_Wb 

where W is the column space of A.  

Notice that b - proj_Wb is in the orthogonal complement of W hence in the null space of A^T.  Hence if x is a this closest vector, then

        A^T(b - Ax)  =  0        A^TAx  =  A^Tb 

Now we need to show that A^TA nonsingular so that we can solve for x.  

Lemma

If A is an m x n matrix of rank n, then A^TA is nonsingular.  

Proof

We want to show that the null space of A^TA is zero.  If 

        0  =  A^TAx

then multiplying both sides by x^T, we get

        0  =  x^TA^TAx  =  (Ax)^TAx  =  Ax ^. Ax  =  ||Ax||²

If the magnitude of a vector is zero, then the vector is zero, hence

        Ax  =  0 

Since 

        rank(A)  =  n

we can conclude that 

        x  =  0 

We can now state the main theorem.

Theorem

Let A be an m x n matrix or rank n, then the system

        Ax  =  b 

has the unique least squares solution 

        x  =  (A^TA) ^-1A^Tb 

Examples

Example

Find the least squares solution to 

        Ax  =  b

with

Solution

We can quickly check that A has rank 2 (the first two rows are not multiples of each other).  Hence we can compute

Notice that 

 not exactly b, but as close as we are going to get.

Least Squares Regression Line

Of fundamental importance in statistical analysis is finding the least squares regression line. 

An engineer is tracking the friction index over mileage of a breaking system of a vehicle.  She expects that the mileage-friction relationship is approximately linear.  She collects five data points that are show in the table below. 

The graph below shows these points

We are interested in the line that best fits the data.  More specifically, if b is the vector of friction index data values and y is the vector consisting of y values when we plug in the mileage data for x and find y by the equation of the line, then we want the line that minimizes the distance between b and y.  If the equation of the line is 

        ax + b  =  y

then we get the five equations

        2a + b  =  20
        6a + b  =  18
        20a + b  =  10
        30a + b  =  6
        40a + b  =  2

The corresponding matrix equation is  

        Ax  =  b 

or

Although this does not have an exact solution, it does have a closest solution.  We have

We can conclude that the equation of the regression line is

        y  =  -0.48x + 20.6

Best Fitting Curves

Often, a line is not the best model for the data.  Fortunately the same technique works if we want to use other nonlinear curves to fit the data.  Here we will explain how to find the least squares cubic.  The process for other polynomials is similar.

Example

A bioengineer is studying the growth of a genetically engineered bacteria culture and suspects that is it approximately follows a cubic model.  He collects six data points listed below

He assumes the equation has the form 

        ax³ + bx² + cx + d  =  y

 This gives six equations with four unknowns

              a +     b +   c + d  =  2.1
            8a +   4b + 2c + d  =  3.5
          27a +   9b + 3c + d  =  4.2
          64a + 16b + 4c + d  =  3.1
        125a + 25b + 5c + d  =  4.4
        216a + 36b + 6c + d  =  6.8

The corresponding matrix equation is 

We can use the least squares equation to find the best solution

So that the best fitting cubic is

        y  =  0.2x³ - 2.0x² + 6.1x - 2.3

The graph is shown below

Back to the Matrices and Vectors Home Page

Back to the Linear Algebra Home Page

Back to the Math Department Home Page

e-mail Questions and Suggestions