PubAg

Main content area

A hybrid model-based framework for estimating ecological risk

Author:
Ni, Lingling, Wang, Dong, Singh, Vijay P., Wu, Jianfeng, Wang, Yuankun, Tao, Yuwei, Liu, Jiufu, Zou, Ying, He, Ruimin
Source:
Journal of cleaner production 2019 v.225 pp. 1230-1240
ISSN:
0959-6526
Subject:
ecosystems, models, risk estimate, support vector machines
Abstract:
For the protection of ecosystems, estimation of ecological risk considering multiple hazards is required. The ecological risk entails subjectivity, randomness, and fuzziness which should be managed appropriately in its estimation. This study developed a framework based on a hybrid model for estimating ecological risk by employing multi-cloud enhanced-fuzzy support vector machine (MC-FSVM). The enhanced FSVM was applied for accounting of multiple hazards, and multi-cloud was used to simultaneously take fuzziness and randomness into account. Based on the prior information, multi-cloud was developed to implement the uncertain transformation between a qualitative concept and its quantitative numerical representation. The degree of certainty derived from the constructed cloud was applied to the enhance fuzzy vector machine to construct an MC-FSVM model. Then by inputting the data into the model trained, ecological risks were graded. This framework was applied to estimate risk in seven regions from five countries and its performance was compared with other two approaches. Results showed the impact of the degree of certainty on the grading of risk and indicated that this framework can be a useful alternative for ecological risk estimation.
Agid:
6367200