Skip to content

M
musicinformationretrieval-com
Commit ef132c53 authored by Leigh Smith's avatar Leigh Smith
Browse files
Options

Upgraded cross-validation and created ToC link to evaluation

parent 14b2f280
Hide whitespace changes
Inline Side-by-side
Showing with 305 additions and 9 deletions
Table_of_Contents.ipynb
View file @ ef132c53
{ {
"metadata": { "metadata": {
"name": "", "name": "",
"signature": "sha256:02701b4647f4627969e1153bf3a0a52f55db3adcfa149e54f81fc4d27d65967b" "signature": "sha256:536c517d48c98583d6306b6a68d4f93eed7dc8277f50f50aa7f338f8a2832ad9"
}, },
"nbformat": 3, "nbformat": 3,
"nbformat_minor": 0, "nbformat_minor": 0,
...@@ -107,14 +107,23 @@ ...@@ -107,14 +107,23 @@
"cell_type": "markdown", "cell_type": "markdown",
"metadata": {}, "metadata": {},
"source": [ "source": [
"1. [Cross Validation](notebooks/cross_validation.ipynb)\n",
"1. [Evaluation](notebooks/evaluation.ipynb)\n",
"1. [Tonal Descriptors: Pitch and Chroma](notebooks/tonal.ipynb)\n", "1. [Tonal Descriptors: Pitch and Chroma](notebooks/tonal.ipynb)\n",
"1. [Feature Extraction](notebooks/feature_extraction.ipynb)\n", "1. [Feature Extraction](notebooks/feature_extraction.ipynb)\n",
"1. [Beat Tracking](notebooks/beat_tracking.ipynb)\n", "1. [Beat Tracking](notebooks/beat_tracking.ipynb)\n",
"1. [Tempo Estimation](notebooks/tempo_estimation.ipynb)\n", "1. [Tempo Estimation](notebooks/tempo_estimation.ipynb)\n",
"1. [Nonnegative Matrix Factorization](notebooks/nmf.ipynb)\n", "1. [Nonnegative Matrix Factorization](notebooks/nmf.ipynb)\n",
"1. [Cross Validation](notebooks/cross_validation.ipynb)\n", "1. [Exercise: Source Separation using NMF](exercises/nmf_source_separation.ipynb)\n"
"1. [Exercise: Source Separation using NMF](exercises/nmf_source_separation.ipynb)"
] ]
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
} }
], ],
"metadata": {} "metadata": {}
......
notebooks/cross_validation.ipynb
View file @ ef132c53
{ {
"metadata": { "metadata": {
"name": "", "name": "",
"signature": "sha256:14bd75857e4b8f9809ba4d5f05dc624688b7b01756fb2f49422023301d45173f" "signature": "sha256:b3f4da5b401fc3911a6fa14e446f660b141ede1e9c80ecd1b470edfb58fc648f"
}, },
"nbformat": 3, "nbformat": 3,
"nbformat_minor": 0, "nbformat_minor": 0,
...@@ -41,6 +41,50 @@ ...@@ -41,6 +41,50 @@
"from sklearn import cross_validation\n", "from sklearn import cross_validation\n",
"from sklearn.neighbors import KNeighborsClassifier\n", "from sklearn.neighbors import KNeighborsClassifier\n",
"from sklearn import preprocessing\n", "from sklearn import preprocessing\n",
"from essentia.standard import MonoLoader\n",
"from essentia.standard import ZeroCrossingRate, CentralMoments, Spectrum, Windowing, Centroid, DistributionShape\n",
"import urllib2\n",
"import urllib\n",
"\n",
"def process_corpus(corpus_URL):\n",
" \"\"\"Read a list of files to process from the text file at corpusURL. Return a list of URLs\"\"\" \n",
" # Open and read each line\n",
" url_list_text_data = urllib2.urlopen(corpus_URL) # it's a file like object and works just like a file\n",
" for file_URL in url_list_text_data: # files are iterable\n",
" yield file_URL.rstrip()\n",
" \n",
"def spectral_features(filelist):\n",
" \"\"\"\n",
" Given a list of files, retrieve them, analyse the first 100mS of each file and return\n",
" a feature table.\n",
" \"\"\"\n",
" number_of_files = len(filelist)\n",
" number_of_features = 5\n",
" features = np.zeros([number_of_files, number_of_features])\n",
" sample_rate = 44100\n",
"\n",
" for file_index, url in enumerate(filelist):\n",
" print url\n",
" urllib.urlretrieve(url, filename='/tmp/localfile.wav')\n",
" audio = MonoLoader(filename = '/tmp/localfile.wav', sampleRate = sample_rate)()\n",
" zcr = ZeroCrossingRate()\n",
" hamming_window = Windowing(type = 'hamming') # we need to window the frame to avoid FFT artifacts.\n",
" spectrum = Spectrum()\n",
" central_moments = CentralMoments()\n",
" distributionshape = DistributionShape()\n",
" spectral_centroid = Centroid()\n",
"\n",
" frame_size = int(round(0.100 * sample_rate)) # 100ms\n",
" # Only do the first frame for now.\n",
" # TODO we should generate values for the entire file, probably by averaging the features.\n",
" current_frame = audio[0 : frame_size]\n",
" features[file_index, 0] = zcr(current_frame)\n",
" spectral_magnitude = spectrum(hamming_window(current_frame))\n",
" centroid = spectral_centroid(spectral_magnitude)\n",
" spectral_moments = distributionshape(central_moments(spectral_magnitude))\n",
" features[file_index, 1] = centroid\n",
" features[file_index, 2:5] = spectral_moments\n",
" return features\n",
"\n", "\n",
"def crossValidateKNN(features, labels):\n", "def crossValidateKNN(features, labels):\n",
" \"\"\"\n", " \"\"\"\n",
...@@ -67,14 +111,14 @@ ...@@ -67,14 +111,14 @@
" \n", " \n",
" # SCALE\n", " # SCALE\n",
" scaler = preprocessing.MinMaxScaler(feature_range = (-1, 1))\n", " scaler = preprocessing.MinMaxScaler(feature_range = (-1, 1))\n",
" trainingFeatures = scaler.fit_transform(features.take(train_index, 0))\n", " training_features = scaler.fit_transform(features.take(train_index, 0))\n",
" # BUILD NEW MODEL - ADD YOUR MODEL BUILDING CODE HERE...\n", " # BUILD NEW MODEL - ADD YOUR MODEL BUILDING CODE HERE...\n",
" model = KNeighborsClassifier(n_neighbors = 3)\n", " model = KNeighborsClassifier(n_neighbors = 3)\n",
" model.fit(trainingFeatures, labels.take(train_index, 0))\n", " model.fit(training_features, labels.take(train_index, 0))\n",
" # RESCALE TEST DATA TO TRAINING SCALE SPACE\n", " # RESCALE TEST DATA TO TRAINING SCALE SPACE\n",
" testingFeatures = scaler.transform(features.take(test_index, 0))\n", " testing_features = scaler.transform(features.take(test_index, 0))\n",
" # EVALUATE WITH TEST DATA - ADD YOUR MODEL EVALUATION CODE HERE\n", " # EVALUATE WITH TEST DATA - ADD YOUR MODEL EVALUATION CODE HERE\n",
" model_output = model.predict(testingFeatures)\n", " model_output = model.predict(testing_features)\n",
" print(\"KNN prediction %s\" % model_output) # Debugging.\n", " print(\"KNN prediction %s\" % model_output) # Debugging.\n",
" # CONVERT labels(test,:) LABELS TO SAME FORMAT TO COMPUTE ERROR \n", " # CONVERT labels(test,:) LABELS TO SAME FORMAT TO COMPUTE ERROR \n",
" labels_test = labels.take(test_index, 0)\n", " labels_test = labels.take(test_index, 0)\n",
...@@ -86,7 +130,250 @@ ...@@ -86,7 +130,250 @@
], ],
"language": "python", "language": "python",
"metadata": {}, "metadata": {},
"outputs": [] "outputs": [],
"prompt_number": 6
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"snares_URL = \"https://ccrma.stanford.edu/workshops/mir2014/SnareCorpus.txt\"\n",
"snare_file_list = [audio_file_URL for audio_file_URL in process_corpus(snares_URL)]\n",
"kicks_URL = \"https://ccrma.stanford.edu/workshops/mir2014/KickCorpus.txt\"\n",
"kick_file_list = [audio_file_URL for audio_file_URL in process_corpus(kicks_URL)]\n",
"all_files = snare_file_list + kick_file_list\n",
"\n",
"drum_labels = np.empty(len(all_files), np.int32)\n",
"num_of_snares = len(snare_file_list)\n",
"drum_labels[0:num_of_snares] = 1 # set labels to the first sample type, e.g. snare\n",
"drum_labels[num_of_snares: num_of_snares + len(kick_file_list)] = 2 # Set next set to the second sample type, e.g kick\n",
"print drum_labels\n",
"\n",
"drum_features = spectral_features(all_files)\n",
"print drum_features\n",
"\n",
"print crossValidateKNN(drum_features, drum_labels)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"[1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2]\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_01_01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_02_01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_04_01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_05_01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_06_01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_07_01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_08_01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_09_01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/SNARE_10_01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/snares/snare_mono.wav"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_01_V01.WAV\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_02_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_03_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_04_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_05_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_06_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_07_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_08_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_09_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"https://ccrma.stanford.edu/workshops/mir2014/audio/drum%20samples/kicks/Bass_Drum_10_V01.WAV"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"[[ 1.29251704e-01 2.20017299e-01 4.07947935e-02 8.77821267e-01\n",
" -8.77261162e-03]\n",
" [ 1.06349207e-01 2.09325567e-01 3.52719985e-02 7.70348966e-01\n",
" -3.22882414e-01]\n",
" [ 1.14965983e-01 1.98087305e-01 3.39526162e-02 8.53771806e-01\n",
" -1.91451073e-01]\n",
" [ 1.09977327e-01 1.89880371e-01 3.07796430e-02 9.61023569e-01\n",
" 2.17356920e-01]\n",
" [ 1.37641728e-01 2.61134952e-01 4.25882451e-02 6.05063379e-01\n",
" -4.61072683e-01]\n",
" [ 1.12925172e-01 2.12166250e-01 3.77560928e-02 9.84243870e-01\n",
" 2.55144835e-01]\n",
" [ 1.43990934e-01 2.32887402e-01 4.25517298e-02 8.88096571e-01\n",
" -2.55122185e-02]\n",
" [ 1.24036282e-01 2.34317511e-01 4.48079035e-02 8.41823637e-01\n",
" -9.77532864e-02]\n",
" [ 1.33560091e-01 2.40808740e-01 3.97951268e-02 7.26550639e-01\n",
" -2.08487749e-01]\n",
" [ 1.14058957e-01 2.19529420e-01 3.08201890e-02 9.37722743e-01\n",
" 6.40640974e-01]\n",
" [ 9.97732393e-03 4.34924886e-02 1.27002802e-02 3.57807994e+00\n",
" 1.42459507e+01]\n",
" [ 7.02947844e-03 2.78101377e-02 6.62755780e-03 4.85479450e+00\n",
" 2.90792427e+01]\n",
" [ 5.66893443e-03 1.62475649e-02 2.56283046e-03 7.87885380e+00\n",
" 8.96873932e+01]\n",
" [ 5.66893443e-03 3.63209695e-02 9.78391711e-03 3.97197366e+00\n",
" 1.83651676e+01]\n",
" [ 8.61677993e-03 2.82980110e-02 8.20268691e-03 5.25167370e+00\n",
" 3.25443840e+01]\n",
" [ 7.93650839e-03 2.40181480e-02 5.99115109e-03 5.35486269e+00\n",
" 3.59657173e+01]\n",
" [ 7.70975044e-03 2.50173863e-02 6.07906608e-03 5.63076735e+00\n",
" 3.99530830e+01]\n",
" [ 5.66893443e-03 3.24563682e-02 9.35256947e-03 4.67505932e+00\n",
" 2.59624958e+01]\n",
" [ 8.84353742e-03 3.49452570e-02 8.03499669e-03 4.01454067e+00\n",
" 2.00484810e+01]\n",
" [ 8.61677993e-03 2.26410441e-02 6.27955049e-03 5.77101851e+00\n",
" 3.90028992e+01]]"
]
},
{
"output_type": "stream",
"stream": "stdout",
"text": [
"\n",
"Fold: 0\n",
"TRAIN: [ 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19]\n",
"TEST: [0 1]\n",
"KNN prediction [1 1]\n",
"Fold: 1\n",
"TRAIN: [ 0 1 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19]\n",
"TEST: [2 3]\n",
"KNN prediction [1 1]\n",
"Fold: 2\n",
"TRAIN: [ 0 1 2 3 6 7 8 9 10 11 12 13 14 15 16 17 18 19]\n",
"TEST: [4 5]\n",
"KNN prediction [1 1]\n",
"Fold: 3\n",
"TRAIN: [ 0 1 2 3 4 5 8 9 10 11 12 13 14 15 16 17 18 19]\n",
"TEST: [6 7]\n",
"KNN prediction [1 1]\n",
"Fold: 4\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 10 11 12 13 14 15 16 17 18 19]\n",
"TEST: [8 9]\n",
"KNN prediction [1 1]\n",
"Fold: 5\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 8 9 12 13 14 15 16 17 18 19]\n",
"TEST: [10 11]\n",
"KNN prediction [2 2]\n",
"Fold: 6\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 8 9 10 11 14 15 16 17 18 19]\n",
"TEST: [12 13]\n",
"KNN prediction [2 2]\n",
"Fold: 7\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 16 17 18 19]\n",
"TEST: [14 15]\n",
"KNN prediction [2 2]\n",
"Fold: 8\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 18 19]\n",
"TEST: [16 17]\n",
"KNN prediction [2 2]\n",
"Fold: 9\n",
"TRAIN: [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17]\n",
"TEST: [18 19]\n",
"KNN prediction [2 2]\n",
"cross validation error: 0.000000\n",
"cross validation accuracy: 1.000000\n",
"None\n"
]
}
],
"prompt_number": 7
}, },
{ {
"cell_type": "markdown", "cell_type": "markdown",
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment