Source

Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree

A basis is a smallest possible set of vectors that still builds the whole space. The number of vectors in it is the dimension.

Basis

A subset of a vector space is a basis for if:

  • is linearly independent

A basis is not unique, e.g. scaling one basis vector by a nonzero still gives a basis. The standard basis of is .

Checking n Vectors Form a Basis of R^n

Put the vectors as columns of a square matrix . Spanning and independence together reduce to one test:

Finding a Basis for span(S)

Two recipes for :

Method 1 (columns)Method 2 (rows)
Put the vectors as columns of Put the vectors as rows of
Reduce to RE formReduce to RE form
Take the original vectors that sit in the pivot columnsTake the nonzero rows of the RE matrix

Method 1 keeps vectors from the original set. Method 2 gives a simpler looking basis.

Worked Example

Find a basis for with .

Pivots are in columns 1, 2 and 4, so a basis is . The span is all of and was redundant.

Dimension

The dimension is the number of vectors in any basis of . Every basis of has the same number of vectors, so dimension is a property of the space itself.

Examples:

  • a line through the origin has dimension 1, a plane through the origin has dimension 2

The most important subspaces to find bases for are those attached to a matrix, see Null Space Column Space and Row Space.