论文标题
使用EEG信号中的时频特征癫痫发作检测和分类
Epileptic Seizure Detection and Classification using Time-Frequency Features in EEG Signals
论文作者
论文摘要
在文献中，使用脑电图信号来诊断几种脑异常。特别是，可以使用EEG信号检测癫痫发作，并在该领域进行了几项工作。事实证明，联合时频域特征是分类结果的改进。这可以归因于EEG信号的非平稳性固有特征。因此，转移到联合时频域在利用频率的时间变化方面受益，反之亦然，以获得更好的结果。贝叶斯分类器用于获得最佳功能的排名。构建信息增益标准是为了测试系统并获得结果。
The use of EEG signal to diagnose several brain abnormalities is well-established in the literature. Particularly, epileptic seizure can be detected using EEG signals and several works were done in this field. The joint time-frequency domain features proved to be an improvement in classification results. This can be attributed to the non-stationarity inherent feature of the EEG signals. Therefore, shifting to the joint time-frequency domain benefits in utilizing the variations in time with respect to frequency and vice versa to obtain better outcomes. The Bayesian classifier is utilized to obtain the ranking of the best features. The information gain criterion was built to test the system and obtain the results.
