论文部分内容阅读
利用BP神经网络对云南高原九大湖泊和长江中下游平原鄱阳湖和巢湖的生命演化阶段进行分类。样本的采集主要包括已知演化阶段的断陷湖和未知演化阶段的断陷湖两部分。根据湖泊演化阶段分类初级指标选取滇池、洱海、抚仙湖(1992~2010年)、鄱阳湖(1994~2012年)、巢湖(1986~1995、2000~2008)18年各指标数据以及程海、泸沽湖、星云湖、异龙湖、杞麓湖、阳宗海(2005~2012年)连续8年数据来进行分类研究。研究结果表明,BP神经网络测试样本能够与训练样本很好的吻合,可以利用其对未知样本分类。分类结果为:青年期湖泊包括抚仙湖、鄱阳湖、泸沽湖;成年期湖泊包括洱海、巢湖、程海、星云湖、杞麓湖、异龙湖和阳宗海;老年期湖泊包括滇池。其中巢湖和异龙湖有向老年期发展的趋势。
BP neural networks were used to classify the stages of the life evolution of nine lakes in Yunnan Plateau and Poyang Lake and Chaohu Lake in the middle and lower reaches of the Yangtze River. The sample collection mainly consists of fault-deprived lakes with known evolution stages and fault-depressed lakes with unknown evolution stages. According to the primary index of lake evolution stage, the indexes of Dianchi Lake, Erhai Lake, Fuxian Lake from 1992 to 2010, Poyang Lake from 1994 to 2012 and Chaohu from 1986 to 1995 and 2000 to 2008, Lugu Lake, Xingyun Lake, Yi Long Lake, Qilu Lake, Yang Zonghai (2005 ~ 2012) data for 8 years to carry out classification studies. The results show that BP neural network test samples can be well matched with the training samples, which can be used to classify unknown samples. The classification results are as follows: The adolescent lakes include Fuxian Lake, Poyang Lake and Lugu Lake; the adult lakes include Erhai Lake, Chaohu Lake, Chenghai Lake, Xingyun Lake, Qilu Lake, Yilong Lake and Yangzonghai Lake; Among them, Chaohu Lake and Yi Long Lake have the tendency to develop toward old age.