Let be an real symmetric matrix with its eigenvalues . Show that for any with , we have
Moreover, when , is an eigenvector of with respect to ; when , is an eigenvector of with respect to .
According to the spectral theorem, a real symmetric matrix is similar to a diagonal matrix by an orthonormal basis. That is, there is an orthonormal basis such that the matrix , when observed by , has its matrix representation diagonal. Intuitively, the proof should be similar to the case when is a diagonal matrix.
By the spectral theorem, there is an orthonormal basis composed of eigenvectors of . We may assume that is the eigenvector corresponding to for .
Since is an orthonormal basis of , we may write
for some scalars . As 's are eigenvectors, we also have
The fact that is orthonormal means that for and for any . By the distributive law of the inner product, we may compute
and
By replacing for , we have
and for . Thus, is the weighted average of with respect to the weights . Consequently, the value of is between the minimum and the maximum of , which are and , respectively. When , necessarily all the weight is attributed to those with , and is a linear combination of eigenvectors with respect to , which is again an eigenvector with respect to . Similarly, when , is an eigenvector with respect to .
This note can be found at Course website > Learning resources.