基于支持向量机的湖泊生态系统健康评价研究

利用支持向量机在处理分类问题、小样本问题和泛化推广方面的优势,构建了基于支持向量机的湖泊生态系统健康评价模型.同时,对广州市最大的人工湖——白云湖的水质及生物群落情况进行了监测,最后运用该模型对白云湖生态系统健康状况进行了评价.评价结果表明,白云湖生态系统处于病态状态,不能达到其净化水质的设计作用.建议从提高进水水质、实施湖区截污和丰富生物量3方面改善白云湖生态系统健康水平.与传统熵权综合健康指数法和熵权模糊综合评价法相比,所建模型更加客观、科学地评价了湖泊生态系统健康状况,能够为湖泊生态系统健康管理提供一定依据,具有广阔的应用前景.

Investigation of health assessment for lake ecosystem based on support vector machine (SVM)

A health assessment model for lake ecosystem was proposed based on the strengths of support vector machine (SVM) on dealing with classification, small sample size, generalization and promotion. Meanwhile, a survey was conducted on the water quality and biological communities of Baiyun Lake, the largest artificial lake in Guangzhou city. Afterwards, the health assessment of Baiyun Lake ecosystem was evaluated with this model. It was demonstrated by the assessment results that the ecosystem of Baiyun Lake was in pathological state, and unable to function in purifying water. To improve the ecological system of Baiyun Lake, three methods were suggested including improving input water quality, cutting pollution sources and enriching biomass. Compared with two traditional evaluation methods (Entropy weight comprehensive health index method and Entropy weight fuzzy comprehensive evaluation method), this model is more objective and scientific on evaluating the health of lake ecosystem. It can provide scientific support for health management of lake ecosystem, therefore with a promising application prospect.