Best Fit:

Normally Engineering data have significant amount of error. Therefore the interpolating curve flowing through all data points is neither necessary nor practical. A curve that gives the best fit to the provided data is what needed in many cases. The best fit curve shows the common trend of the data (without passing through every point). Various criteria might be used to derive the best fit curve. This is very useful in characterizing the data or in deriving the correlating equation. For instance, correlating equations demonstrating the behaviour of I.C. engines under several air-fuel ratios are useful in the design and analysis of automobiles.