The document discusses the problem with interpolating polynomials and introduces splines as an alternative approach. Splines divide the interpolation interval into smaller sections and fit lower order polynomials within each section rather than a single high order polynomial over the entire interval. This allows for greater control of the interpolating function between data points. Specifically, the document covers:
- Interpolating polynomials lack control between data points
- Splines divide the interval into sections and fit separate polynomials (e.g. lines or parabolas) in each section
- Quadratic splines use parabolas in each section, joined at the endpoints with continuous slopes
- The spline coefficients are determined by satisfying the constraints at endpoints and joining slopes between