What question did this study set out to answer?

This research aims to forecast skill demand and predict salaries in the evolving labor market using advanced machine learning techniques.

May 3, 2026Open Access

Skill Demand Forecasting and Salary Prediction: A Multi-Granularity Analysis Using XGBoost

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial forecasts skill demand and salary in various markets, suggesting improved prediction methods.

This research aims to forecast skill demand and predict salaries in the evolving labor market using advanced machine learning techniques.
Analyzed real job postings from 2021 to 2023, constructing a dataset with millions of rows.
Applied XGBoost with engineered features, including lagged values and rolling averages, for demand forecasting.
Evaluated model performance separately on months with zero demand and conducted feature importance and clustering analyses.
Achieved a symmetric mean absolute percentage error (SMAPE) of 10.01% for predicting active demand.
Generated an R² of 0.164 for salary predictions using various job-related attributes.
Found that a rolling average is the strongest predictor for skill demand, and occupation-level forecasts are most accurate.

Md Zahidul Islam Sany (2026) studied this question.

View Full Paper

Ask AI

Bookmark

May 3, 2026Open Access

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial forecasts skill demand and salary in various markets, suggesting improved prediction methods.

This research aims to forecast skill demand and predict salaries in the evolving labor market using advanced machine learning techniques.
Analyzed real job postings from 2021 to 2023, constructing a dataset with millions of rows.
Applied XGBoost with engineered features, including lagged values and rolling averages, for demand forecasting.
Evaluated model performance separately on months with zero demand and conducted feature importance and clustering analyses.
Achieved a symmetric mean absolute percentage error (SMAPE) of 10.01% for predicting active demand.
Generated an R² of 0.164 for salary predictions using various job-related attributes.
Found that a rolling average is the strongest predictor for skill demand, and occupation-level forecasts are most accurate.

Md Zahidul Islam Sany (2026) studied this question.

View Full Paper

Ask AI

Bookmark