【作者】 郭嘉良；

【导师】 王洪礼；

【作者基本信息】 天津大学 ， 管理科学与工程， 2010， 博士

【摘要】 海岸带自然资源丰富、地理位置突出并具有独特的特性,已成为人类社会经济生产活动活跃度最高的地区。我国海岸带也同样是社会经济发展最迅速的区域,但与此同时,海岸带生态区域又承受着极大的环境生态压力。随着社会经济的快速发展,对资源的不断索取以及污染物排放等压力也日益严峻,造成了海岸带区域原本脆弱敏感的生态环境更加不堪重负。因此,迫切需要对海岸带生态经济系统中经济发展与环境保护之间的关系进行深入研究。本文使用经济与环境发展中的库兹涅兹曲线理论、数理统计中的非参数统计方法-随机梯度回归算法、集成学习技术中的规则集成方法,并结合定性分析与定量计算,对海岸带渔业生态经济系统中经济发展和环境质量之间的关系进行了深入研究与预测,主要取得了以下成果:首先,选定包括近岸海域水体质量以及空气质量等描述海岸带区域环境质量的指标和描述渔业经济增长的指标,再利用环境库兹涅兹曲线(EKC)理论,建立海岸带城市经济-环境计量模型,还利用天津市实际数据进行了曲线拟合,具体分析了天津市EKC曲线的演变机制。并在保证经济增长的前提下,为改善海岸带生态经济系统的环境质量提供了政策建议。其次,在自然-环境-社会三者相互作用过程中,分析了影响海岸带渔业产值的主要因素;根据压力-状态-响应(P-S-R)模型,建立海岸带生态经济系统指标体系;首次采用随机梯度回归法,建立了海岸带生态经济系统中渔业产值的压力胁迫模型,并分析了主要胁迫因素。利用天津市最近十九年的实际数据,通过胁迫模型,得到了天津市渔业产值的各主要影响因素及各因素的影响程度。最后,利用集成学习方法,在本文建立的海岸带生态经济系统指标基础上,选择对渔业产值影响较大的因素,利用ARMA-GARCH模型,建立了基于规则集成学习法的预测模型,对天津市渔业经济的发展进行了实证预测并对相应的实际经济意义进行了分析。结果表明,规则集成法比支持向量机法及自适应样条回归等方法的拟合效果更好,具有更高的预测精度。更多还原

【Abstract】 Coastal zone is rich in resources, with the features of high productivity, great natural conditions, convenient geographic position and special property characteristic. Coastal zone has become the most economically invigorating region during human social and economic production activities. Coastal zone eco-economic system is very important. In China, coastal zone regions are the most rapidly growing economic regions. Because the economy develops fast and demand of resource is getting more and more, there are serious pressures on the coastal zone eco-economic system, such as large amount land-sourced pollutant emissions and unreasonable fishery. Therefore, it is necessary to do research on the relationship between the economic development and protection of environment.In this dissertation, coastal zone eco-economic system was analyzed by the Environmental Kuznets Curve Theory, non-parametric statistical methods such as stochastic gradient regression algorithm and ensemble learning methods such as RULE ENSEMBLE method. Main results and conclusions are shown as follows: First, according to Environmental Kuznets Curve (EKC) theory, measurement models of the relationship between economy development and environment quantity of the coastal zone city were established based on selected indicators of water quality, air quality and economic growth in coastal zone. Then the characters of the EKC were analyzed by the actual data of Tianjin. And policy recommendations were analyzed to ensure economic growth and improve coastal environmental quality eco-economic system provides policy recommendations.Second, the main factors of coastal fishery production were analyzed according to the nature-environment-social interaction. Using the pressure - state - response (P-S-R) model, a coastal eco-economic system index system was established. It is for the first time using stochastic gradient regression method to establish fisheries production model with the influence of stress factors, then the influence of each factor was analyzed. Using the actual data of Tianjin in recent years, there are analysis of the fishery output value of Tianjin and the major factors that influenced the degree. Finally, using the ensemble learning methods, the prediction model of fisheries production was established based on the ecological and economic system index, ARMA-GARCH model and RULE ENSEMBLE learning method. Then do some research on the development of Tianjin Fisheries production and the actual economic significance of the corresponding analysis. The results show that the RULE ENSEMBLE method performs well then support vector machine method and the adaptive regression spline method with better accuracy.更多还原