Chord-Recognition-Using-Machine-Learning

File Usage

• params.yaml
  • parameters setting
• utils.py
  • user defined function / class
• preprocess.py
  • transfer .wav files into a json file using pitch class profiler (pcp)
  • mode = "train" / "test" (set in params.yaml)
  • python preprocess.py
• KNN.py, SVM.py, DecisionTree.py
  • input: train json file
  • algorithm implementation
  • python [KNN/SVM/DecisionTree].py
• compare.py
  • imput: train json file
  • compare performance between algorithms
  • python compare.py
• test.py
  • input: test json file
  • get prediction result of given file
  • python test.py

Dataset Hierarchy

.
├── NewDataset
│   ├── A
│   ├── Am
│   ├── Bb
│   ├── Bdim
│   ├── Bm
│   ├── C
│   ├── D
│   ├── Dm
│   ├── E
│   ├── Em
│   ├── F
│   ├── G
│   └── data.json
├── TestWave
│   ├── about_a_girl.wav
│   ├── test.json
│   └── wave
└── params.yaml