静态异速生长对布氏田鼠和棕色田鼠头骨识别模型的影响

Effect of Static Allometric Growth to Brandt's Voles and Brown Rats Skull Recognition Model

目前异速生长定律多应用于植物的生长和动物的代谢,但在鼠类分类领域尚未有报道,为了研究静态异速生长对头骨识别模型的影响,选取成年组的布氏田鼠和棕色田鼠的头骨,用BugshapeV 1.0获取44个数学形态特征,对其中的变量特征进行比值化处理,非变量特征不做处理,共得到40个数学形态比值特征,对原始特征和比值特征及进行分析建模,用于鼠种的分类识别。通过对比两类模型的判别率可以看出,两类模型的判别率一致。这说明静态异速生长对建立的模型判别率影响很小,可直接用原始数据进行分析建模,这可能是因为分析软件会筛选出更具分辨能力的特征进行建模。

Allometry most used in studies on plant growth and animal metabolism. And it has not been reported in the field of rodent classification. In order to study the influence of Static allometry on the skull identification models. The paper selected adult skulls of Lasiopodomys brandti and Lasiopodomys mandarinus as samples, and got 44 features with Bugshape V1.0. After variables features ratio processing, two kinds of models were built, ratio features model and raw features model. The results show that models built by ratio features and raw features were same accuracy in initial, cross and applied discriminant on species. This indicated static allometry has weak influence on modeling, and raw features can be used to model.