Context: Systematic literature reviews (SLRs) are the primary method for aggregating and synthesizing evidence in evidence-based software engineering. Primary study selection is a critical and time-consuming SLR step in which reviewers use titles, abstracts, or even full texts to evaluate thousands of studies to find the dozens of them that are relevant to the research questions. Objective: We seek to reduce the effort of primary study selection in SLRs with machine assisted reading techniques. Method: In this paper we explore and refactor the state-of-the-art machine assisted reading techniques from both evidence-based medicine and legal electronic discovery to support SLRs. By refactoring those methods, we discovered FASTREAD, which is a new state-of-the-art in machine assisted primary studies for SLRs. Tested on two data sets generated from existing SLRs of Hall, Wahono, et al., FASTREAD outperforms the current state-of-the-art methods. Results: Using FASTREAD, it is possible to find 90% of the studies found by standard manual methods, but after only reading less than 10% of the candidate studies. Conclusions: With the help of FASTREAD, conducting an SLR is much more efficient and less difficult. Software Engineering researchers now have no excuse: they should conduct SLRs.
from cs.AI updates on arXiv.org http://ift.tt/2hgG4ZR
via IFTTT
No comments:
Post a Comment