文件列表:

B_matrix.mat (354, 2017-12-16)
Hilbert_Envelope.m (1915, 2017-12-16)
Homomorphic_Envelope_with_Hilbert.m (3501, 2017-12-16)
backward_prop.m (1321, 2017-12-16)
backward_prop2.m (1518, 2017-12-16)
butterworth_high_pass_filter.m (2906, 2017-12-16)
butterworth_low_pass_filter.m (2910, 2017-12-16)
classes_lrhsmm.csv (1555, 2017-12-16)
classes_train.csv (9055, 2017-12-16)
classes_val.csv (752, 2017-12-16)
cross_entropy.m (389, 2017-12-16)
default_Springer_HSMM_options.m (1646, 2017-12-16)
error_analysis_LRHSMM.m (4872, 2017-12-16)
evaluate_metric.m (1051, 2017-12-16)
expand_qt.m (2231, 2017-12-16)
extractFeaturesFromHsIntervals.m (6487, 2017-12-16)
features_gt.csv (82378, 2017-12-16)
features_lrhsmm.csv (83582, 2017-12-16)
features_train.csv (447154, 2017-12-16)
features_val.csv (41489, 2017-12-16)
forward_prop.m (1165, 2017-12-16)
forward_prop2.m (1511, 2017-12-16)
getDWT.m (2523, 2017-12-16)
getHeartRateSchmidt.m (3781, 2017-12-16)
getSpringerPCGFeatures.m (4179, 2017-12-16)
get_PSD_feature_Springer_HMM.m (2514, 2017-12-16)
get_duration_distributions.m (3838, 2017-12-16)
k_means.m (122656, 2017-12-16)
labelPCGStates.m (7091, 2017-12-16)
lr_model.m (7454, 2017-12-16)
main.m (8917, 2017-12-16)
main2.m (9156, 2017-12-16)
normalise_signal.m (1457, 2017-12-16)
pca.m (795, 2017-12-16)
pi_vector.mat (188, 2017-12-16)
runSpringerSegmentationAlgorithm.m (2742, 2017-12-16)
run_LR_HSMM.m (7715, 2017-12-16)
schmidt_spike_removal.m (4383, 2017-12-16)
... ...

# CS229_Project_HeartSound Application of Machine Learning Techniques for Heart Sound Recording Classification By Vincent Lee and Anatoly Yakovlev Dataset and LR-HSMM starter code were provided by Physionet: https://www.physionet.org/challenge/2016/ The dataset is not included in this repo. Please download them separately from https://www.physionet.org/physiobank/database/challenge/2016/ Also, example_data.mat training data for LR-HSMM can be downloaded from https://www.physionet.org/physiotools/hss/ To run only classification on supplied features follow these steps: 1. For logistic regression run: lr_model.m 2. For weighted logistic regression run: weighted_LR.m 3. For K-means clustering classification run: k_means.m 4. For SVM classification run: svm_train.m 5. For 3-layer neural network run: main.m 6. For 4-layer neural network run: main2.m To run full flow, first download all the datasets mentioned above and follow these steps: 1. To segment data and extract features from all training datatsets run: run_LR_HSMM.m This step will loop for validation, train-a, -b, -c, -d, -e, -f folders in ../ directory and save all extracted features from train directories into features_train.csv, labels for training examples are saved in classes_train.csv. It will save all extracted features from validation examples into features_val.csv, labels for validation examples are saved in classes_val.csv. 2. To run classification on extracted features follow classification steps. Principal component analysis (PCA) can be run to reduce dimensionality of features. To run classification modes on transformed data: 1. Run pca.m. This will save features_train_pca.csv and features_val_pca.csv 2. All the classification models neet to set pca flag to 1 (or True). The following are files provided by Physionet: butterworth_high_pass_filter.m butterworth_low_pass_filter.m default_Springer_HSMM_options.m expand_qt.m getDWT.m get_duration_distributions.m getHeartRateSchmidt.m get_PSD_feature_Springer_HMM.m getSpringerPCGFeatures.m Hilbert_Envelope.m Homomorphic_Envelope_with_Hilbert.m labelPCGStates.m runSpringerSegmentationAlgorithm.m schmidt_spike_removal.m trainBandPiMatricesSpringer.m trainSpringerSegmentationAlgorithm.m viterbiDecodePCG_Springer.m viterbi_Springer.c viterbi_Springer.mexa*** The following are files provided by Physionet, but modified by us: extractFeaturesFromHsIntervals.m The following are files created by us: (Neural Network Models: NN1, NN2) backward_prop2.m backward_prop.m cross_entropy.m sigmoid_func.m forward_prop2.m softmax_func.m forward_prop.m main.m main2.m (Logistic Regression) lr_model.m (Locally Weighted Logistic Regression) weighted_LR.m (K-means) k_means.m (Gaussian kernel SVM) svm_train.m svm_test.m (PCA for feature selection) pca.m (Data Segmentation and Feature Extraction) run_LR_HSMM.m normalise_signal.m (Error / Performance Analysis) error_analysis_LRHSMM.m evaluate_metric.m The following are data files: (LR-HSMM trained parameters) B_matrix.mat pi_vector.mat total_obs_distribution.mat (LR-HSMM training dataset) example_data.mat (Extract features) features_lrhsmm.csv features_train.csv features_val.csv features_gt.csv (Data labels) classes_lrhsmm.csv classes_val.csv classes_train.csv

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：