The well-known equation for the evaluation of relative growth rate (RGR) RGR=ln(w(2)/w(1))/(t(2)-t(1)) was shown to be an average relative rate. Only in the case of exponential growth law is this equation valid for both average and instantaneous rates. It depends not only on the time interval of averaging [t(1), t(2)] but also on the error of measurements. Under certain conditions, real growth rate could not be seen among noise resulted from data errors. Other equations often used such as RGR=[w(2)-w(1)]/[w(1)((t(2)-t(1))] Or its modification RGR=[w(1)+w(2)]/[0.5((w(1)+w(2))((t(2)-t(1))] and %increase={(w(2)/w(1))boolean AND[1/ (t(2)-t(1))]-1}100%, do not describe rates and cannot be used for the purpose in mind. Such a situation impelled us to develop a new approach for determining instantaneous rates directly from the experimental data for any process. The idea of this method consists of an approximation of data by a cubic spline regression having first and second derivatives. The analytical differentiation of the spline regression permits the determination of instantaneous rate. The method of minimisation of the functional of average risk was used successfully to solve the problem. This method permits to obtain the instantaneous rate directly from the experimental data. The instantaneous rate is a highly sensitive characteristic for study of natural and anthropogenic influences on the biological and ecological processes, For illustration, we analysed heat production of microplankton and growth rate of red seaweed Gracilaria verrucosa versus temperature. The program is written in C++.