Test of Randomness of Residuals for Modified Gompertz Model used for Modelling the Growth of Callus Cultures from Glycine wightii (Wight & Arn.) Verdc.

Abstract

One of the most important preliminary investigations of callus attributes is the growth characteristics. Most often than not, callus growth curve is sigmoidal in characteristics. Frequently, plant scientists studying callus growth neglect the utilization of mathematical growth that are useful in obtaining important growth constants such as lag period, maximum specific growth rate and maximum growth or asymptote. Formerly, we model callus growth of Glycine wightii from published literature to obtain vital growth constants. We discovered that the
modified Gompertz model via nonlinear regression utilizing the least square method was the best to explain the growth curve. Nevertheless, an important thing to consider, that has not been stated more than enough, is the residual of the model needs to be random. To make sure that randomness being fulfilled we carry out the Wald-Wolfowitz runs test. The results demonstrated that the number of runs was 5, and the expected number of runs within the assumption of randomness was 5, suggesting the series of residuals had perfect runs. The p-value obtained was higher than 0.05, hence the null hypothesis is not rejected suggesting no persuading proof of nonrandomness
of the residuals plus they do stand for noise.