23.07.2012 Views

Linear Algebra

Linear Algebra

Linear Algebra

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Topic: Line of Best Fit 269<br />

Topic: Line of Best Fit<br />

This Topic requires the formulas from the subsections on Orthogonal Projection<br />

IntoaLine,andProjectionIntoaSubspace.<br />

Scientists are often presented with a system that has no solution and they<br />

must find an answer anyway, that is, they must find a value that is as close as<br />

possible to being an answer. An often-encountered example is in finding a line<br />

that, as closely as possible, passes through experimental data.<br />

For instance, suppose that we have a coin to flip, and want to know: is it<br />

fair? This question means that a coin has some proportion m of heads to flips,<br />

determined by how it is balanced beween the two sides, and we want to know<br />

if m =1/2. We can get experimental information about it by flipping the coin<br />

many times. This is the result a penny experiment, including some intermediate<br />

numbers.<br />

number of flips 30 60 90<br />

number of heads 16 34 51<br />

Naturally, because of randomness, the exact proportion is not found with this<br />

sample — indeed, there is no solution to this system.<br />

30m =16<br />

60m =34<br />

90m =51<br />

That is, the vector of experimental data is not in the subspace of solutions.<br />

⎛ ⎞ ⎛ ⎞<br />

16 30<br />

⎝34⎠<br />

�∈ {m ⎝60⎠<br />

51 90<br />

� � m ∈ R}<br />

However, as described above, we expect that there is an m that nearly works.<br />

An orthogonal projection of the data vector into the line subspace gives our best<br />

guess at m.<br />

⎛ ⎞ ⎛ ⎞<br />

16 30<br />

⎝34⎠<br />

⎝60⎠<br />

⎛ ⎞<br />

51 90<br />

30<br />

⎛ ⎞ ⎛ ⎞ · ⎝60⎠<br />

=<br />

30 30<br />

90<br />

⎝60⎠<br />

⎝60⎠<br />

90 90<br />

7110<br />

12600 ·<br />

⎛ ⎞<br />

30<br />

⎝60⎠<br />

90<br />

The estimate (m = 7110/12600 ≈ 0.56) is higher than 1/2, but not by much, so<br />

probably the penny is fair enough for flipping purposes.<br />

The line with the slope m ≈ 0.56 is called the line of best fit for this data.<br />

heads<br />

60<br />

30<br />

��<br />

��<br />

��<br />

30 60 90<br />

flips

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!