Neural Network for ECG signal
in [Matlab]
|
Prev: DENDROGRAM GUIDANCE
Next: Matlab code for kernel -pca
From: prashant Narvekar on 17 Sep 2005 10:56 hi how can i used matlab to develop neural network for ecg signal classification. please guide me. regards
From: Peter K. on 17 Sep 2005 20:24 prashant Narvekar <pmn_18(a)yahoo.co.in> writes: > how can i used matlab to develop neural network for ecg signal > classification. Use your natural neural network. Or buy: http://www.mathworks.com/products/neuralnet/ Ciao, Peter K.
From: prashant Narvekar on 17 Sep 2005 22:05 Dear Sir, thanks for the link. i m novice to matlab and student of Artificial Intelligence. i want to develop neural network system for ecg signal classification as part of artificial intelligence subject. how do i start and where i will get the required infomation for this project to begin. how do i program the matlab for that to start. i told to use som for classification.how i will write matfile for that. please help me. regards prashant.
From: Peter K. on 18 Sep 2005 11:56 prashant Narvekar <pmn_18(a)yahoo.co.in> writes: > thanks for the link. > i m novice to matlab and student of Artificial Intelligence. > i want to develop neural network system for ecg signal classification as part of artificial intelligence subject. > how do i start and where i will get the required infomation for this project to begin. > how do i program the matlab for that to start. > i told to use som for classification.how i will write matfile for that. > please help me. Well, what form do you hav tgeh ECG signal in? Is that in matlab? What sort of classification do you want? Is the ECG data that you have labelled according to the way you want to classify it? If you have ECG signals that are labelled, for example, "Good" and "Bad", then you _might_ be able to use this directly to train a learning algorihtm... though I doubt it. More often, you have to analyse the data and see what "features" are best to discriminate between the "Good" and the "Bad". This usually means some sort of feature extraction has to be done. Even if you use the data directly, you might have to time-align all the data to use it sensibly. Ciao, Peter K.
From: Greg Heath on 19 Sep 2005 06:14
prashant Narvekar wrote: > Dear Sir, > thanks for the link. > i m novice to matlab and student of Artificial Intelligence. > i want to develop neural network system for ecg signal classification as part of artificial intelligence subject. > how do i start and where i will get the required infomation for this project to begin. > how do i program the matlab for that to start. > i told to use som for classification. Kohonen's SOM is an unsupervised learning algorithm that was meant for visualization. It was not meant for classification. Kohonen used SOM to initialize a sequence of supervised learning classification algorithms called LVQ1, LVQ2, LVQ3, etc. However, I don't recommend LVQ over other supervised learning algorithms like Levenberg-Marquardt or Conjugate Gradient which are contained in the Neural Net Toolbox. Hope this helps. Greg > how i will write matfile for that. > please help me. > regards > prashant. |