WRDScore: New Metric for Evaluation of Natural Language Generation Models

Evaluating natural language generation models, particularly for method name prediction, poses significant challenges. A robust metric must account for the versatility of method naming, considering both semantic and syntactic variations. Traditional overlap-based metrics fail to capture these nuances. Existing embedding-based metrics often suffer from imbalanced precision and recall, lack normalized scores, or make unrealistic assumptions about sequences. To address these limitations, we propose WRDScore, a novel metric that strikes a balance between simplicity and effectiveness. Our metric is lightweight, normalized, and precision-recall-oriented, avoiding unrealistic assumptions while aligning well with human judgments.