An ambient air quality evaluation model based on improved evidence theory.

Area of Science:

• Environmental Science
• Data Science
• Decision Theory

Background:

• Scientific air quality evaluation is crucial for managing air pollution.
• Dempster-Shafer (D-S) evidence theory can address uncertainties in air quality assessment.
• Existing research on D-S theory for air quality assessment is limited, and the standard combination rule can yield counterintuitive results.

Purpose of the Study:

• To propose an improved evidence theory, the DCre-Weight method, to address limitations in D-S theory for comprehensive decision-making.
• To develop and validate an air quality evaluation model (DCreWeight model) based on the improved evidence theory.
• To enhance the credibility and accuracy of air quality comprehensive assessment.

Main Methods:

• Developed the DCre-Weight method, incorporating evidence weights determined by the entropy weight method and decision credibility calculated from evidence dispersion.
• Proposed the DCreWeight model for air quality evaluation.
• Validated the model using hourly air pollution data from Xi'an (June 2014 - May 2016) and compared it against D-S theory and other evaluation methods.

Main Results:

• The DCre-Weight method demonstrated improved fusion result credibility and better uncertainty expression in algorithm cases.
• The DCreWeight model showed superior performance compared to D-S theory, other improved evidence theory methods, and a fuzzy synthetic evaluation method.
• Under the national Air Quality Composite Index (AQCI) standard, the DCreWeight model achieved a Mean Absolute Error (MAE) of 1.02 and Root Mean Square Error (RMSE) of 1.17.
• Under the national Air Quality Index (AQI) standard, the DCreWeight model exhibited minimal MAE and RMSE, and the maximal index of agreement.

Conclusions:

• The DCreWeight model effectively and comprehensively evaluates air quality, offering a scientific basis for air pollution control.
• The improved evidence theory (DCre-Weight method) provides reasonable fusion results and overcomes limitations of the standard D-S theory.
• The study validates the superiority of the DCreWeight model in air quality assessment, highlighting its potential for practical application.