Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2004

Experimental data came in discrete form in real world situations. These points may be interpolated to approximate the function. However, observational errors are always associated with experimental data. This study compared the performances of four piecewise interpolation methods, namely, piecewise linear, piecewise quadratic, a modified piecewise cubic polynomial and cubic spline in interpolating data points with random errors on four selected test functions. Random errors were generated based on two distribution functions (i.e. Gaussian distribution with parameters and and uniform distribution with range [a, b]). Piecewise linear performed best over the other hand, piecewise quadratic interpolation is better on small values of and smaller ranges. On oscillatory type of function, cubic spline performed best on small values of and smaller ranges. Moreover, the modified cubic polynomial performed best on very small values of and ranges used in generating random errors. As an illustration, the recommended method was applied to the data of monthly solid waste disposal of Davao City for CY 1996-2002 based solely on the collection of City Environment and Natural Resources Office (CENRO).