Research on Object-Oriented Programming Teaching Evaluation Method Based on Few-Shot Prompt Learning

Abstract

This paper proposes a programming education evaluation method based on few-shot prompt learning for object-oriented programming education. The method consists of three core components: establishing multi-dimensional evaluation criteria, designing a few-shot learning-based example code library, and developing intelligent evaluation prompt templates. Through a 16-week controlled experiment conducted in a university computer science program, results demonstrate significant improvements in both learning outcomes and teaching efficiency: the experimental group showed a 9.2% to 15.4% improvement in key metrics including final exam scores and code quality, while reducing grading time by 80% and shortening feedback time from 48 hours to 0.5 hours. This research provides a new paradigm for educational evaluation in computer science education.