Research on optimization of enterprise salary structure based on deep learning algorithm

In order to optimize the enterprise salary structure and improve the scientificity and fairness of salary distribution, deep learning algorithm is used to model and optimize the salary structure. The deep neural network (DNN) model is constructed by analyzing more than 5,000 employee data and combining multi-dimensional data such as employee performance, job level and market salary. The experimental results show that the deep neural network performs well in predicting the optimization adjustment of salary, with the mean square error (MSE) of the test set being 0.0187 and the coefficient of determination (R²) being 0.943. The optimized salary structure improves the employee satisfaction by an average of 10.7%, and saves the enterprise's salary cost by 5.56%. The analysis concludes that the deep learning algorithm can effectively improve the accuracy and rationality of compensation management.