نوع مقاله : مقاله پژوهشی
نویسندگان
1 بخش تحقیقات علوم دامی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی یزد، سازمان تحقیقات، آموزش و ترویج کشاورزی، یزد، ایران
2 بخش تحقیقات علوم دامی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان قم، سازمان تحقیقات، آموزش و ترویج کشاورزی، قم، ایران
3 استادیار بخش تحقیقات علوم دامی، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی گلستان، سازمان تحقیقات، آموزش و ترویج کشاورزی، گرگان، ایران
چکیده
کلیدواژه‌ها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Background and objectives: Camel breeding is one of the sources of income for people on the edge of arid and desert areas in many parts of the world. Traits related to camel growth, including birth weight, weaning weight, daily weight gain and one-year-old weight are considered as the main economic traits for camel owners. To manage and improve the genetic and phenotypic values of these traits, weight recording of animals is essential. Recording of camels, especially their weight is associated with many difficulties due to their restless temperament and large size. Using efficient mathematical methods can greatly solve this problem. Various approved efficient mathematical methods have been proposed to predict the weight of camels based on their body dimension, and their effectiveness has been proven. So, the present study was conducted to compare the efficiency of the Principal Component Analysis and multiple regressions in estimating the weight of fattening camels from its body dimensions.
Materials and methods: In order to compare the efficiency of principal components analysis and multiple regressions in estimating the weight of fattening camels based on body dimensions, the records of 220 fattening camels of Bafgh station in Yazd were used. For this purpose, new born camels were fed for a 9 months period using standard diets. During the period, each of the fattening camels was weighed and different body sizes including body length(BL), whither height(WH), breast girth(BG), abdomen with(AW), hump height to the ground(HH), muzzle girth(MG), neck length(NEL), whither to pin length(WPL), tail length(TL), pelvic width (PW), abdomen to hump height (ABH) and the head length(HL) were measured. The body dimensions of the camels were recorded using a tape measure and their body weight was recorded using a scale. Then, the data were analyzed using the principal component analysis and multiple regressions. In order to fit the predictive models, body weight of camels was introduced as dependent variable and body size of camels as independent variables. Analysis of regression models was done using one and multivariable linear models and the best model were selected to estimate the weight of camels based on their body dimensions by Stepwise method. The performance of the above models was evaluated using comparison of the coefficient of determination (R2) of them.
Results: According to the results, the correlation between the body weight of camels with their different body dimensions including BL, WH, BG, AW, HH, MG, NEL, WPL, TL, PW,ABH and HL were 0.93, 0.89, 0.89, 0.89, 0.94, 0.73, 0.89, 0.90, 0.80, 0.85, 0.89 and 0.79 ,respectively. The results showed that among the 6 multiple regression models fitted to estimate the weight of fattening camels, model No. 6 in which head length, body length, breast girth, neck length, muzzle girth and whither height were used as predictive variables had the least error (12.06) and the highest accuracy (0.92) compared to other models. The results showed that the use of the first and second principal component , and both of them in the model could explain 82.1%, 3.73% and 85% of the variance of body weight, respectively. The weight of fattening camels was determined using principal component analysis with an accuracy of 0.93 and an error of 11.54.
Conclusion: The results of the present study showed that in order to estimate the weight of fattening camels, the use of principal component analysis, in addition to simplifying predictive models, has higher efficiency and accuracy as well as less error compared to multiple regressions, and this method can be a suitable alternative to the multiple regression method in predicting the weight of camels.
کلیدواژه‌ها [English]
)