Computer program for fitting the logistic and the Gompertz function to growth data: From a study concerning the adolescent growth cycle in men

✍ Scribed by E. Marubini; L.F. Resele

• Publisher: Elsevier Science
• Year: 1971
• Weight: 362 KB
• Volume: 2
• Category: Article
• ISSN: 0010-468X
• DOI: 10.1016/0010-468x(71)90004-3

✦ Synopsis

In the study of human growth, particular emphasis is placed on the period from pre-puberty to maturity. For the interpolation of data collected by means of both longitudinal and cross-sectional types of surveys, growth experts utilise two asymptotic-type functions: the logistic and the Gompertz function.

However, the use of either is difficult, for widescale applications. This is because it is necessary to fit to the experimental points non-linear functions, owing to which the straightforward method of multiple linear regression, based on the minimum squares principle, cannot be used.

The authors therefore, in order to obtain the minimum variance estimates of the constants applied the least squares principle after linearization in implementing the computing program in FOTRAN IV language.

The program can either calculate initial estimates of the constants, or utilise estimates known a priori to the researcher, and the iterative process is continued until the decrease of residual sum of squares between two successive iterations is less than a suitably selected value.