论文部分内容阅读
据《Scientia Horticulturae》的一篇研究报道(http://dx.doi.org/10.1016/j.scienta.2013.12.012),来自巴西的J.D.R.Soares等人比较了几种香蕉产量估算技术。田间试验观察到的植物技术特点均是表型特征,其估算大部分是以观察者的经验为依据。变量相关性分析可以在其他特性变化的基础上进行某种特征变化估计。研究人员研究了利用栽培特性预测产量的潜能,在香蕉估产中应用了两项变量相关性分析技术:人造神经网络(ANNs)和多元线性回归(MLR)。这项试验是对两种分析估产方法一致性的检测。供试品种为Tropical(YB42-21),
According to a study by Scientia Horticulturae (http://dx.doi.org/10.1016/j.scienta.2013.12.012) J.D.R.Soares et al. From Brazil compared several techniques for estimating banana yields. The technical characteristics of plants observed in field trials are all phenotypic characteristics, and the majority of their estimates are based on the observer’s experience. Correlation analysis of variables can make some kind of feature change estimation based on the changes of other features. Researchers studied the potential for predicting yield using cultivation traits. Two variable correlation analysis techniques, artificial neural networks (ANNs) and multiple linear regression (MLR), were applied to banana yield estimation. This test is a test of the consistency of the two analytical methods. Test varieties for the Tropical (YB42-21),