Output details
11 - Computer Science and Informatics
Lancaster University
Learning from examples to improve code completion systems
<08> The first work to use machine learning on big data extracted by static analysis of software repositories to enable intelligent code completion. Recommends only frequently observed API usage patterns in a specific context, thus significantly increasing the efficiency of software development. The work initiated the Eclipse CodeRecommenders open-source project, now a popular feature of the Eclipse distribution, which is used by more than 6 million Java developers. The work was published in a leading software engineering conference (ESEC/FSE), acceptance rate 32/217 (15%) in 2009, and has found great attention in the scientific community as reflected by the citation rate.