An Automatic Approach to Validating Log Levels in Java

  • Taeyoung Kim
  • , Suntae Kim
  • , Cheol Jung Yoo
  • , Soohwan Cho
  • , Sooyong Park

    Research output: Contribution to conferenceConference paperpeer-review

    Abstract

    A log statement is used to record important runtime behavior of software systems for diverse reasons, which is inevitable to develop most of the software systems. However, developers do not tend to deeply consider an appropriate log level in their source code. In order to address the issues, this paper proposes an automatic approach to validating log levels in Java in consideration of the syntactic as well as semantic features. We first build up the Word2Vec model and generate semantic and syntactic log feature vectors, then train the machine learning classifiers to automatically validate the log levels. For the evaluation, we collected six open source projects of the message-oriented middleware domain, and obtained the 88% precision and the 87% recall respectively.

    Original languageEnglish
    Title of host publicationProceedings - 25th Asia-Pacific Software Engineering Conference, APSEC 2018
    PublisherIEEE Computer Society
    Pages623-627
    Number of pages5
    ISBN (Electronic)9781728119700
    DOIs
    StatePublished - 2018.07.2
    Event25th Asia-Pacific Software Engineering Conference, APSEC 2018 - Nara, Japan
    Duration: 2018.12.42018.12.7

    Publication series

    NameProceedings - Asia-Pacific Software Engineering Conference, APSEC
    Volume2018-December
    ISSN (Print)1530-1362

    Conference

    Conference25th Asia-Pacific Software Engineering Conference, APSEC 2018
    Country/TerritoryJapan
    CityNara
    Period18.12.418.12.7

    Keywords

    • classification model
    • log feature vector
    • log level
    • log statement
    • Word2Vec

    Quacquarelli Symonds(QS) Subject Topics

    • Computer Science & Information Systems

    Fingerprint

    Dive into the research topics of 'An Automatic Approach to Validating Log Levels in Java'. Together they form a unique fingerprint.

    Cite this