论文部分内容阅读
敏感性分析能够定量地评价模型输入变量的变化对输出结果产生的影响,是揭示模型蕴含规律的有效途径.本文将敏感分析方法应用于BP神经网络巢湖水华预测模型中,分析结果表明巢湖水华形成受诸多环境因子共同影响,水温、溶解氧、浊度、气温、光照强度等环境因子变化与藻类质量浓度变化相关,其中气温是最大影响因素,相对贡献率达到17.01%;气压的上升则不利用于藻类质量浓度的增加;pH值的上升对藻类质量浓度的影响有正有负.
Sensitivity analysis can quantitatively evaluate the impact of the change of model input variables on the output results, which is an effective way to reveal the law of implicated model.In this paper, the sensitivity analysis method is applied to the BP neural network model of Chaohu Lake bloom forecasting, the results show that Chaohu water The formation of China was influenced by many environmental factors. The changes of environmental factors such as water temperature, dissolved oxygen, turbidity, temperature and light intensity were related to the change of algae mass concentration, of which the temperature was the most important factor and the relative contribution rate reached 17.01% Not conducive to the increase in the concentration of algae; pH value of the algae have a positive and negative impact on the concentration.