| Autor: |
Norbert Brunner, Manfred Kühleitner |
| Jazyk: |
angličtina |
| Rok vydání: |
2020 |
| Předmět: |
|
| Zdroj: |
Veterinary and Animal Science, Vol 10, Iss , Pp 100135- (2020) |
| Druh dokumentu: |
article |
| ISSN: |
2451-943X |
| DOI: |
10.1016/j.vas.2020.100135 |
| Popis: |
Growth literature often uses the Brody, Gompertz, Verhulst, and von Bertalanffy models. Is there a rationale for the preference of these classical named models? The versatile five-parameter Bertalanffy-Pütter (BP) model generalizes these models. We revisited peer-reviewed publications from the years 1970–2019 that fitted growth models to together 122 mass-at-age data of sheep and goats from 19 countries and studied the best-fit BP-models using the least-squares method. None of the named models was ever best-fitting. However, for 70% of the data a single non-sigmoidal model had an acceptable fit (normalized root mean squared error 〈 5% and F-ratio test 〉 5% in comparison to the best-fit): the Brody model. The inherently non-sigmoidal character was further underlined, as there were only 39% of the data, where the best-fitting BP-model had a discernible inflection point. For these data, conclusions of biological interest could be drawn from the sigmoidal best-fit BP-models: the maximal weight gain per day was about 55% higher than the natal weight gain per day. |
| Databáze: |
Directory of Open Access Journals |
| Externí odkaz: |
|
