Comparing logit and probit logistic regression in classification of treatment on Heros severous based on hight and weight

Estimating the feed treatment groups of ornamental fish, Heros severus, based on weight and length gain through the rearing time, is the aim of the present study. 300 gold severum with initial weight and length of 0/5±0/02 g and 2/2±0/25 cm, respectively were divided into three treatments, each with six replicates (TA: Specimens feed with diet contain fish oil; TB: Specimens feed with diet contain soybean oil; TC: specimens feed with diet contain equal shares of fish oil and soybean oil). Feeding trail with Iso-nitrogenous diets lasts for 90 days and every fifteen days fish biometry were being practiced. Regard to the classification of the three treatments, practical models including “cumulative logit” and “cumulative probiot” were used. Moreover through “Backward elimination” the best models were obtained and being compared. Logistic model with logit link function being introduced as the best result.