Predictive Salary Modeling for IT-specialists

The article presents an analysis of data on the remuneration of specialists from organizations in different countries in order to select significant features to create a regression model that predicts the level of wages based on such factors as the position held, work experience, type of employment, form of work (remote, hybrid, in the office), and company size. To select the most optimal regression model, three machine learning methods were used: decision tree, linear regression, gradient boosting. The calculated RMSE quality metrics showed the need to exclude the decision tree model, due to the presence of a significant error in its predictions. The remaining two models were used to test the salary prediction for a candidate for the position of Lead Data Management Specialist with randomly generated data on experience and working conditions.