Skip to content

Minutes

Mikhail Veselov edited this page Mar 26, 2016 · 4 revisions

February, 26, 2016

  • Project introduction: problem formalization, organization, future plans
  • Two variants for the result:
  • recommendation script
  • browser extension
  • Library choosing
  • Share knowledge (article pdfs, site links, etc).
  • Futher plans for the next week:
  • All possible data sources
  • Library analysis
  • Article search and investigation
  • How to measure the project quality? (WTF)

March, 4, 2016

  • We should concentrate on feature extraction, not on learning method, as there are lots of working approaches
  • What features are used in articles we've reviewed? Create a feature list page with comments
  • How do we get & store the music libraries?
  • Last.Fm API limitations investigation (Mikhail)
  • Librosa histograms extraction (Natalia)
  • Essentia histograms extraction (Mikhail)

March, 11, 2016

  • Text extraction should be more automatical, some samples for further investigation
  • Still no data, unfortunately
  • Split content-based calculation: Librosa vs Essentia
  • Get similar information from Last.FM
  • Get ID3 tags from mp3 file
  • Store data into JSON

March, 21, 2016

  • Meeting time changed
  • Still no data, unfortunately
  • We need much more code and data :)
  • Texts gather
  • LDA for tags for songs
  • Dataset exchange
Clone this wiki locally