Predicting the Salary in Data Science through BP Neural Network

With the development of technology, working in data science is becoming more and more popular for its high salary. Therefore, forecasting salary has become a more concerning issue. This paper focuses on the fundamental principles, structure, and algorithm of the BP neural network for salary prediction. The feasibility of establishing BP neural network model with MATLAB was trained and tested with the data of the relevant variable data from the average salary of data engineers from the Kaggle website. The results show that the predicated salary fits well with the true value through BP neural network. Moreover, the relative error of the model is around 10% and the duration is 2s, which is quite acceptable. Besides, in order to know the results more directly, 20 samples are shown in the report, which contains the highest relative error is -16.89%, while the lowest one is zero. Finally, based on the accuracy of the model in predicting the future average salary, it can provide some reference value for job seekers.