A novel evaluation model for heavy-metals pollution in soil based on connection numbers and Dempster–Shafer theory

The evaluation of heavy-metals pollution in soil involves with various uncertainty factors. Although the normal cloud model provides an idea for dealing with the randomness and fuzziness of indicators for the evaluation of heavy-metals pollution in soil, it distorts data fusion and cannot simulate the distribution state of evaluation indicators in a finite interval. Herein, a novel cloud model coupled with connection numbers and Dempster–Shafer evidence theory was discussed to remedy the defects of the normal cloud model. In the model proposed here, the connection cloud model was first presented to simulate the classification standard for each evaluation indicator. Then, the connection degree was calculated to construct the evaluation matrix, and the corresponding basic belief assignment matrix was obtained on the basis of the Dempster–Shafer evidence theory. Next, combining together with the combination weights gained by a distance function and the consistency between the differences in the static and dynamic weights and their coefficients, the classification of heavy-metals pollution in soil was determined according to the mean evidence value. Finally, the case study and comparisons with the fuzzy mathematical method, the extension method, the risk index method and the improved grey clustering method were conducted, respectively, to confirm the validity and reliability of the proposed model. Results show that this novel model can overcome the shortcomings of the normal cloud model and provide a method for comprehensive evaluation of heavy-metals pollution in soil.