Author(s)

Mengqiu Kong

Corresponding Author

Mengqiu Kong

ABSTRACT

This paper theoretically proves that the least square method for linear fitting after logarithmic transformation of data is essentially based on the principle of reducing relative error. Through empirical analysis and comparison with the traditional least square method, it is found that the least square method after logarithmic transformation of data has a higher precision in model fitting effect, and at the same time makes it possible to use the least square method after logarithmic transformation of data to fit the model accurately. Fitting the straight line gives better consideration to the information of all observation points.

KEYWORDS

Linearity after logarithmic, transformation, least square method