shiridisaibaba22/Bias_Detection-Fairness_Evaluation_in_Hiring_Demo

⚖️ Bias Detection & Fairness Evaluation in Hiring Algorithms Goal: Use the UCI Adult Income dataset as a proxy for hiring decisions (predict whether income >50K) and:

Train a classification model (Logistic Regression) for hiring decisions. Evaluate using traditional metrics (accuracy, F1) and fairness metrics (demographic parity, equal opportunity, equalized odds). Analyze bias across gender (sex) and race, and demonstrate two simple mitigation strategies: reweighing (pre-processing) and group-specific thresholding (post-processing). Dataset (UCI Adult / "Census Income"):

Data: https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data Attribute information: https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.names Note: In this notebook we treat income >50K as the model's positive outcome (hiring/promotion/selection proxy).