Using Logistic Regression and Ensemble Learning for Employment Status Prediction

As digital technology reshapes the labor market, real-time insights into employment dynamics have become key to the smooth operation of the economy. Based on 5,000 anonymized sampling data in Yichang City, this study constructs a multi-dimensional employment status analysis and prediction framework. The first step is to use data cleaning and visualization to evaluate the influence of age, gender, education and other characteristics to reveal structural characteristics such as high youth unemployment rate and weak female employment stability. In the second step, the chi-square test was used to screen significant variables, and a logistic regression model was constructed to predict the employment status of 20 test samples, with an accuracy of 81.93% and a recall rate of 97.7%. The third step is to introduce macroeconomic indicators such as GDP growth rate, urban registered unemployment rate, and policy support level, and use the random forest model to optimize the prediction after integrating with individual data, with an accuracy of 81.2% and an F1 value of 0.90, confirming the moderating effect of macro factors on employment. This study provides a data-driven decision-making basis for regional employment policy formulation and targeted assistance.