The instrument measurement in hedging boiler has hysteresis

which causes the control system always lag behind the controlled object. Therefore

in order to understand the trend of NO_x at outlet of the hedge boiler in advance with changes in working conditions

it is necessary to quickly and accurately predict the NO_x at outlet of the hedge boiler. Based on the actual operating unit data

this paper uses Random Forest algorithm(random forest)to establish a 1 000 MW hedged boiler unit outlet NO_x prediction model

and uses Spearman coefficient to filter the features of the hedged boiler to further reduce the model calculation time. The results show that

the root mean square error of the prediction results of the random forest model is 10.182 mg/m~3

and the coefficient of determination is 0.913. Therefore

the model structure of the random forest based on feature selection can realize the forecast of the NO_x trend at outlet of the hedging boiler in advance.