Target Variable

A target variable, in machine learning and related fields, represents the outcome or dependent variable being predicted or analyzed. Current research focuses on improving the definition and handling of target variables, particularly addressing challenges like existential quantification in AI planning (using graph neural networks) and the impact of target variable definition on fairness in machine learning models (using interactive simulators). These efforts aim to enhance the accuracy, efficiency, and ethical implications of models by optimizing target variable engineering, including exploring both regression and classification approaches and extending quantile regression to handle vector-valued targets with improved computational efficiency.