Previous Lecture

lect08

Next Lecture

lect08, Tue 02/04

Data fitting by least squares with QR factorization

• Lecture Files

Reading assignment

For Thursday, read Section 2.9 (norms and condition numbers) of the NCM book.

References for today’s lecture

Sections 5.1 through 5.5 of the NCM book.

Outline

• Manipulating sparse matrices in numpy/scipy

• Data fitting by least squares using QR factorization
  • The factorization A = QR for non-square matrices A
  • Least squares problems: Ax $\approx$ b
    • a surveying problem
  • Solving Ax $\approx$ b by QR factorization
• numpy/scipy routines:
  • scipy.sparse.csr_matrix()
  • spla.qr() [note: this is slightly different from npla.qr]
  • npla.lstsq()