Salary Prediction of AI Jobs Using Machine Learning

Abstract

In the modern world, salary serves as a crucial motivator for employees, making accurate salary prediction significant for both employers and employees. It enables both parties to estimate expected compensation effectively, facilitating better career planning, resource allocation, and informed negotiations. With advancements in Data Science and Machine Learning, predicting salaries has become increasingly viable and reliable. This study leverages a dataset of over 65,000 salaries from the Stack Overflow Annual Developer Survey and explores four supervised machine learning techniques: Linear Regression, Decision Tree, Random Forest, and Tuned Random Forest. These models are evaluated using performance metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R² to determine their accuracy and predictive capabilities. The Tuned Random Forest model demonstrated the best performance with an RMSE of 23,426 and an R² score of 0.53, achieving higher accuracy than other models. These findings highlight the impact of hyperparameter tuning in enhancing model effectiveness and confirm the Tuned Random Forest as a reliable tool for salary prediction in AI-related job roles. This research underscores the contribution of machine learning techniques to salary decision-making, with potential applications in other sectors of the industry.