Exact Sciences Department, Sao Paulo State University, Jaboticabal, Brazil.
Poult Sci. 2012 Oct;91(10):2710-7. doi: 10.3382/ps.2011-01878.
Experimental studies have shown that hatching rate depends, among other factors, on the main physical characteristics of the eggs. The physical parameters used in our work were egg weight, eggshell thickness, egg sphericity, and yolk per albumen ratio. The relationships of these parameters in the incubation process were modeled by Fuzzy logic. The rules of the Fuzzy modeling were based on the analysis of the physical characteristics of the hatching eggs and the respective hatching rate using a commercial hatchery by applying a trapezoidal membership function into the modeling process. The implementations were performed in software. Aiming to compare the Fuzzy with a statistical modeling, the same data obtained in the commercial hatchery were analyzed using multiple linear regression. The estimated parameters of multiple linear regressions were based on a backward selection procedure. The results showed that the determination coefficient and the mean square error were higher using the Fuzzy method when compared with the statistical modeling. Furthermore, the predicted hatchability rates by Fuzzy Logic agreed with hatching rates obtained in the commercial hatchery.
实验研究表明，孵化率取决于蛋的主要物理特性等因素。我们在工作中使用的物理参数是蛋重、蛋壳厚度、蛋球形度和蛋黄与蛋白比例。这些参数在孵化过程中的关系通过模糊逻辑进行建模。模糊建模的规则基于孵化蛋的物理特性以及使用商业孵化场分析和各自的孵化率，通过在建模过程中应用梯形隶属函数来实现。实现是在软件中完成的。为了比较模糊逻辑与统计建模，使用商业孵化场获得的相同数据，通过多元线性回归进行分析。多元线性回归的估计参数基于后向选择过程。结果表明，与统计建模相比，模糊方法的确定系数和均方误差更高。此外，模糊逻辑预测的孵化率与商业孵化场获得的孵化率一致。
