Data smoothing based on local procedures
Guerra Ortiz, Victor Manuel
Tapia, Richard A.
Master of Arts
This thesis presents a method for smoothing data based on local procedures. The method attempts to reproduce what is actually done in manual loft. Properties of this method are that points with large errors are smoothed very quickly, an that large quantities of data can be smoothed without the traditional problems of excessive computer time and numerical instability which are usually associated with non-local procedures, We offer two approaches for the smoothing method, the first is based on linear interpolation and is studied in great detail? the second is based on cubic interpolation. In the first approach only lines are completely smooth while in the second only cubics are completely smooth.