The Application Of Rough Set And Artificial Neural Network To Electrocardiogram Auto-analyzing | Posted on:2003-02-25 | Degree:Master | Type:Thesis | Country:China | Candidate:J Qi | Full Text:PDF | GTID:2144360065961182 | Subject:Operational Research and Cybernetics | Abstract/Summary: | PDF Full Text Request | This paper proposes the improvement of LADT , So it can be realized faster than before. The role of rough set is studied in the feature appreciation and selection and the equivalence condition of the roughness in knowledge is obtained and proved. So the RS theory is enriched. We attain features and choose from them by use of above LADT method and RS theory. With above features we try two methods to recognize R waves: one is the way of auto-recognizing R waves directly with the constant threshold of the feature; another is the method that R wave are recognized by the trained Neural Network whose input data is the value of various feature. The result shows that the second one is more valid. In the end we use neural network combined with RS theory to realize ventricular QRS classification, thus we set up the basis for the following ECG auto-diagnosing.
| Keywords/Search Tags: | LADT, method, rough set, ECG, Featuer extraction feature selection, QRS wave, R wave, Neural network ventricular wave | PDF Full Text Request | Related items |
| |
|