Research on Employee Feedback and Strategy Formulation Based on Natural Language Processing and Deep Learning

In this paper, the extreme value gradient growth method of Xgboost is used to grade and screen customers by using customer value probability. Assist staff to prejudge customer categories. Daily statistics and intuitive tips on the impact of personnel workload and customer conversion are provided through automatic operation procedures and Tableau reports. In this way, the feedback of personnel performance reaches the closed loop of the whole system. Utilize UML to craft the architecture of the system, followed by employing the visual blueprint to present the system through intuitive visuals, encompassing use scenario chart, classification diagram, interaction layout, and behavioral diagram. The system's use scenario model, construction model, and behavioral model are crafted visually. After the system completes more systematic data storage management, more accurate customer partitioning, and timelier closed-loop monitoring, it is obvious that high-quality differential operation can be achieved. After the system was put into use, the various conversions of users increased by 1.5-10 times compared to the initial stage.