Empirical Evaluation of ChatGPT on Requirements Information Retrieval Under Zero-Shot Setting

TL;DR

This paper empirically assesses ChatGPT's effectiveness in requirements information retrieval tasks under zero-shot conditions, highlighting its strengths in recall and limitations in precision for specific requirements.

Contribution

It introduces an evaluation framework for ChatGPT on requirements IR tasks and provides initial insights into its capabilities and limitations in this domain.

Findings

High recall in retrieving relevant requirements

Limited precision in retrieving specific requirements

Provides evidence for future IR method development using LLMs

Abstract

Recently, various illustrative examples have shown the impressive ability of generative large language models (LLMs) to perform NLP related tasks. ChatGPT undoubtedly is the most representative model. We empirically evaluate ChatGPT's performance on requirements information retrieval (IR) tasks to derive insights into designing or developing more effective requirements retrieval methods or tools based on generative LLMs. We design an evaluation framework considering four different combinations of two popular IR tasks and two common artifact types. Under zero-shot setting, evaluation results reveal ChatGPT's promising ability to retrieve requirements relevant information (high recall) and limited ability to retrieve more specific requirements information (low precision). Our evaluation of ChatGPT on requirements IR under zero-shot setting provides preliminary evidence for designing or developing more effective requirements IR methods or tools based on LLMs.