Coke quality prediction is meaningful to the coal blending and coking quality control of coke production facilities.In this study,recent two-years coal quality data of a coking plant is used,combining with relative parame-ters of production process,to build prediction model for several coke quality targets.Both coking theory and feature engineering methods are used to select features,along with prediction results of multiple machine learning models compared on the dataset,to determine the most predictive model.The result shows that,the model is predictive of the coke quality of the coking plant.
coke qualityfeature engineeringmachine learningpredictive model
维普
万方数据
