Existing approaches to recover links between e-mails and software artifacts are based on text search or text retrieval and reformulate link recovery as a document retrieval problem. We refine and improve such solutions by leveraging the parts of which an e-mail is composed of: header, current message, and previous messages. The relevance of these parts is weighted by a probabilistic approach based on text retrieval. We implemented our novel solution exploiting the BM25F model. The results of an empirical study conducted on a public benchmark indicate that the new approach in many cases outperforms the baseline approaches chosen. In addition, the proposed approach is easy to use and it is accurate enough to be worth the costs it may introduce in the corpus preprocessing and indexing.
Linking E-Mails and Source Code Using BM25F
G. Scanniello
2013-01-01
Abstract
Existing approaches to recover links between e-mails and software artifacts are based on text search or text retrieval and reformulate link recovery as a document retrieval problem. We refine and improve such solutions by leveraging the parts of which an e-mail is composed of: header, current message, and previous messages. The relevance of these parts is weighted by a probabilistic approach based on text retrieval. We implemented our novel solution exploiting the BM25F model. The results of an empirical study conducted on a public benchmark indicate that the new approach in many cases outperforms the baseline approaches chosen. In addition, the proposed approach is easy to use and it is accurate enough to be worth the costs it may introduce in the corpus preprocessing and indexing.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
