Detect type of change in a project analyzing the log history
a project by aplanas
Updated about 3 years ago. 8 hacker ♥️.

Use machine learning and natural language processing techniques to analyze the changes made in a project, and classify them in:

  • Small / unimportant fix
  • Big / important fix
  • Small / important feature
  • Big / important feature

For this project I will

  1. Generate a basic corpus of labeled data from a different set of project related with openSUSE
  2. Evaluate the best features to make a proper classification: n-gram, PoS tag, TF-IDF (with and without stemmer)
  3. Evaluate and measure the best classification model: Naive Bayes, Linear SVM, Max Entropy, ...

Join this project

Looking for hackers with the skills:

nlp machinelearning git github

This project is part of:

Hack Week 10 Hack Week 11 Hack Week 12

Activity

  • over 8 years ago: jordimassaguerpla liked this project.
  • almost 9 years ago: nicolasbock liked this project.
  • about 10 years ago: ZRen disliked this project.
  • about 10 years ago: ZRen liked this project.
  • about 10 years ago: bkutil liked this project.
  • over 10 years ago: cschum liked this project.
  • over 10 years ago: froh joined this project.
  • over 10 years ago: vitezslav_cizek liked this project.
  • over 10 years ago: froh liked this project.
  • over 10 years ago: oholecek liked this project.
  • over 11 years ago: aplanas liked this project.
  • over 11 years ago: aplanas started this project.
  • over 11 years ago: aplanas added keyword "nlp" to this project.
  • over 11 years ago: aplanas added keyword "machinelearning" to this project.
  • over 11 years ago: aplanas added keyword "git" to this project.
  • over 11 years ago: aplanas added keyword "github" to this project.
  • over 11 years ago: aplanas originated this project.


    • Comments

    • aplanas
      over 10 years ago by aplanas | Reply

      Yeah. Hackweek 10 collied with openSUSE 13.1, so I will try to for on this during this new Hackweek instance : )

    • froh
      over 10 years ago by froh | Reply

      Would it be hard to train for regression fix vs new feature, based on the comment? I'd be curious how much energy project have to put into regression fixes vs feature additions.

    • osynge
      about 10 years ago by osynge | Reply

      Have you considered looking at ELK and integrating this work in the ELK stack.

    Similar Projects

    This project is one of its kind!