Aiming at the characters and problems of soft sensor, a soft sensor modeling method based on ensemble learning
algorithm AdaBoost RT is presented. According to the shortages of AdaBoost RT and the difficulties of updating for soft
sensor model, the self-adaptive modified value of ?? and the incremental learning character added improved methods are
proposed. The method is used to establish the soft sensor model of molten steel temperature in 300 t LF. The product data
are used to test the model. The test results show that the proposed soft sensor model based on improved AdaBoost RT can
improve the prediction accuracy and update real time.