Listing of MATLAB Programs

• Steps for running the program:
  1. Change suvParamSet.m, if necessary, to change the parameters for SUV detection.
  2. goCollectData.m
     • Collect the training/test data.
     • Store the data at DS.mat.
  3. goSelectInput.m
     • Load DS.mat for input selection using LOO of KNNR.
     • Save the indices of the selected inputs at bestInputIndex.mat.
  4. goPlotData2d.m: Plot the 2d scatter data of the selected inputs.
  5. goPlotDataAll.m: Plot all combinations of the inputs of the data.
  6. goQcDesign.m
     • Design a quadratic classifier.
     • Save the parameters as qcParam.mat.
     • Plot the decision boundary if the feature dimension is 2.
  7. goGmmcDesign.m
     • Find the no. of mixtures for each class of GMM.
     • Save the parameters as gmmData.mat.

• List of M files:

  File Name	H1 Help	File Size
  decBoundaryPlot.m	This script is called by goDesignClassifier.m and goTest.m	765
  getRecogRate.m	getRecogRate: Compute the recognition rate of SU/V detection	1383
  goDataCollect.m	Collect the data of MIR-1K	760
  goGmmcDesign.m	Classifier design based on GMM	1552
  goHmmTestAll.m	====== Load parameters	2856
  goHmmTestOne.m	load bestInputIndex.mat	979
  goHmmTrainTestAll.m		184
  goInputSelect.m	Input selection based on KNNC and LOO	1802
  goPlotDataAll.m	2D/3D scatter plots for data visualization	963
  goPlotSelectedInput.m	2D scatter plot	677
  goQcDesign.m	Design a quadratic classifier	1284
  goTest.m	Test the classifier for SU/V detection	496
  goTransProb.m		727
  goTutorial.m	% Tutorial on Voiced Sound Detection for Singing & Humming Recordings	9052
  hummingFeaExtract.m	hummingFeaExtract: Wave to feature conversion for s/u/v detection	4742
  myClassifier.m	myClassifier: My classifier for SU/V detection	246
  vsdHmmTrainTest.m	% === Collect wave data	661
  vsdOptSet.m	vsdOptSet: Set options for VSD (voiced sound detection) for singing/humming	882
  waveDecode.m	load bestInputIndex.mat	2092