基于因子分析-BP神经网络模型在空气质量综合评价中的应用

发布时间：2018-03-17 21:27

本文选题：BP神经网络　切入点：因子分析　出处：《云南大学》2015年硕士论文　论文类型：学位论文

【摘要】：综合评价在社会经济各个领域有着广泛应用,经过多年的研究,逐渐形成了多种评价方法,每种方法有各自的优点和缺陷,因此,一直处于不断改进和完善的过程中。近些年来,随着网络的广泛应用和发展,数据信息量日益增大,对综合评价方法的发展提出了新的要求,因此对多指标综合评价方法的研究更加有实际意义。近年来比较流行的人工神经网络法,以其独特的分布式处理能力、非线性处理能力、自学功能和容错性等成为了解决复杂且难以建模问题的强有力工具。通过对数据的多次训练,在输出结果中直接对变量进行等级(类别)判定。在一定程度上弥补了以前许多方法过于依赖主观判断或客观特定数据的不足,成为目前被广泛应用的一种综合评价方法。当指标过多时,为了提高网络的学习能力,加快处理问题的速度以及增加评价结果的可靠性,需要对网络模型进行简化,可以配合统计方法对多指标进行有效降维,这样可以缩小神经网络模型在处理较多指标时的结果偏差,使评价结果更加可靠。将因子分析引入BP神经网络模型进行BP神经网络综合评价方法的改进研究是本论文的主旨,具体的思路是先采用因子分析法将多指标进行降维,再运用BP神经网络模型进行综合评价,将该模型应用到环境空气质量的实证评价中,以国家2012年环境空气质量新标准中提出的AQI法为评价参考,经过试验,验证了因子分析-BP神经网络相融合的多指标综合评价模型在环境空气质量评价中具有实用性和可行性。同时,通过与单纯使用神经网络模型作对比,进一步验证了该方法的有效性。并且,该模型还可以进一步拓展应用到其他领域的类似多指标综合评价研究中。
[Abstract]:Comprehensive evaluation has been widely used in various fields of social economy. After many years of research, various evaluation methods have been gradually formed, each method has its own advantages and disadvantages, therefore, In recent years, with the extensive application and development of the network, the amount of data information is increasing day by day, which puts forward new requirements for the development of comprehensive evaluation method. Therefore, the study of multi-index comprehensive evaluation method has more practical significance. In recent years, artificial neural network (Ann), which is popular in recent years, has its unique distributed processing ability and nonlinear processing ability. Self-learning and fault tolerance have become powerful tools for solving complex and difficult modeling problems. To a certain extent, it makes up for the deficiency of many previous methods that rely too much on subjective judgment or objective and specific data. In order to improve the learning ability of network, accelerate the speed of processing problems and increase the reliability of evaluation results, the network model should be simplified. It is possible to reduce the dimension of multiple indexes effectively with statistical methods, which can reduce the deviation of the results of the neural network model in dealing with more than one index, and make the evaluation results more reliable. It is the main idea of this paper to introduce factor analysis into BP neural network model to improve the comprehensive evaluation method of BP neural network. Then the BP neural network model is used for comprehensive evaluation, and the model is applied to the empirical evaluation of ambient air quality. The AQI method proposed in the new national standards for ambient air quality in 2012 is used as a reference for the evaluation. The feasibility and practicability of the multi-index comprehensive evaluation model based on factor analysis and BP neural network fusion in the assessment of ambient air quality are verified. At the same time, the model is compared with the neural network model. The validity of the proposed method is further verified. Furthermore, the model can be further extended to other fields of similar multi-index comprehensive evaluation research.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X823

【参考文献】