Abnormal analysis of moisture content of tobacco leaves by big data based on XGboost model

In order to solve the problem of traditional modeling cannot solve the problem of unstable moisture content at the outlet of Moisture machine, Select material flow rate, water addition ratio, circulating hot air temperature during operation, compensation fresh air temperature, circulation damper opening, compensation for fresh damper opening, discharge hood pressure, outlet tobacco temperature, outlet tobacco moisture content, steam application ratio as Process parameters; then the process parameter correlation analysis model and the water content anomaly marking model were determined, and 69 abnormal rules of moisture content of tobacco sheets and 120 normal rules were excavated by XGboost based on big data. The accuracy and recall rate of the model were 92.5% and 96.7%, respectively. The uniformity and accuracy of the moisture improves the stability of the moisture content of the tobacco sheet after regaining moisture.