Fig. 6: Importance of input features for 20 models, including a combined model and 9 sector-independent models to predict total generation quantity of HW, 5 classification models and 5 regression models from the combined and independent ensemble model to predict MHW generation quantity. | Nature Communications

Fig. 6: Importance of input features for 20 models, including a combined model and 9 sector-independent models to predict total generation quantity of HW, 5 classification models and 5 regression models from the combined and independent ensemble model to predict MHW generation quantity.

From: Data-driven approaches linking wastewater and source estimation hazardous waste for environmental management

Fig. 6

a The average relative importance, indicated by the average MAS value of variables in 4 groups of the firm scale, manufacturing processes, the wastewater routine monitoring indicators, and metal emission in wastewater. b Ranking of input feature importance for 20 models. The number inside each cell reflected the ranking of the variable. Variables ranked in the top 5 are marked with *** to denote their crucial importance in model prediction, while variables ranked between 6th and 10th are marked with * to indicate their relatively significant impact.

Back to article page