Further Examples: MatrixRank

Here is a matrix.

In[1]:=

Out[1]//MatrixForm=

The rank of a matrix of exact numbers is computed by row reduction.

In[2]:=

Out[2]=

The rank of a matrix of inexact numbers can be computed by taking the length of the list of positive singular values of the matrix.

In[3]:=

Out[3]=

Using the option Tolerance gives a different result.

In[4]:=

Out[4]=

The difference is due to the algorithm, which generates roundoff errors. Usually Mathematica replaces small numbers of order $MachineEpsilon with unless Tolerance is set to .

In[5]:=

Out[5]=

In[6]:=

Out[6]=

In[7]:=

Out[7]=

Sometimes using higher precision will help to correctly determine the rank.

In[8]:=

Out[8]=

The rank of a matrix is the difference between the dimension of the range and the dimension of the null space.

In[9]:=

Out[9]=

MatrixRank works with symbolic matrices.

In[10]:=

Out[10]=

MatrixRank also works with complex matrices.

In[11]:=

Out[11]=

You can set Tolerance to be large.

In[12]:=

Out[12]=

In[13]:=

Out[13]=

In[14]:=