Representing Documents

• Natural language representations inadequate

• Use a Vector Space representation

dog comput window ...

Doc. 1 = < 1, 0, 1, ... >

Doc. 2 = < 0, 1, 1, ... >

Doc. 3 = < 0, 1, 0, ... >

• Text can have 103 to 106 dimensions!

5

Previous slide

Next slide

Back to first slide

View graphic version