mikhailamod/bn_one

BN Assignment1

All the data fields:

• Genre (has 16 genres)
• Readability, discretized into:
  • x<3
  • 3<=x<5
  • 5<=x<7
  • x>=7
• Difficulty, discretized into:
  • x<4
  • 4<=x<5
  • 5<=x<7
  • x>=7
• Number of duplicates, discretized into:
  • x<20
  • 20<=x<30
  • 30<=x<50
  • x>=50
• Number of words in song, discretized into:
  • x<200
  • 200<=x<300
  • 300<=x<400
  • x>=400
• Sentiment, discretized into:
  • Negative: x<-0.05
  • Neutral: -0.05<=x<0.05
  • Positive: x>=0.05
• WonGrammy (boolean)

This also has a split_data function which takes 20 random songs from each year and produces a testdata.json, and puts the rest into traindata.json