In this study, we worked on EEG data which is taken from the patients with epilepsy and healthy volunteer people. Data consist of healthy signals and unhealthy signals belongs to 500 subject that are labeled as 5 different class. Instead of using classical signal processing techniques, we tried to implement machine learning techniques and develop a system that can be used to diagnose epilepsy. K- nearest neigborbood algorithm and naïve bayes algorithms are used and compared. As result of our experiment, k-nearest neighborhood algorithm successful than naïve bayes. However, remarkable result is gathered after princible component analysis (PCA) to reduce dimension of dataset. In general, we expect to see increased prediction success graphic after PCA but for this medical data set PCA gave the negative results. In addition to this, decreasing the number of labels while analyzing the results, showed more successful prediction result and impact of the illnesses to different area of the brain. You can find the experimential results and detailed information about data set in this article.
[1]
Enas W. Abdulhay,et al.
Automated diagnosis of epilepsy from EEG signals using ensemble learning approach
,
2017,
Pattern Recognit. Lett..
[2]
Miseon Shim,et al.
Machine-learning-based diagnosis of schizophrenia using combined sensor-level and source-level EEG features
,
2016,
Schizophrenia Research.
[3]
K Lehnertz,et al.
Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: dependence on recording region and brain state.
,
2001,
Physical review. E, Statistical, nonlinear, and soft matter physics.
[4]
Isaac Fernández-Varela,et al.
Combining machine learning models for the automatic detection of EEG arousals
,
2017,
Neurocomputing.