Kubra Eroglu

Fast and accurate PLS-based classification of EEG sleep using single channel data

Fast classification of sleep and wake stages using a single EEG channel is proposed.The dataset was provided by Physionet.Speed and accuracy of PLS were compared with those of k-NN, Bayes and LDC classifiers.Results indicated that the Pz-Cz channel had better accuracy than the Fpz-Cz channel.We achieved 91% classification accuracy by selecting PLS as the classifier. Since speed of classification is important to real-time applications, this study proposed fast classification of sleep and wake stages using a single electroencephalograph (EEG) channel. Changes in the sleep and wake stages are accompanied by changes in the frequency spectrum of the EEG signals; so, the features extracted from the 5-s epoch of the EEG using auto-regressive (AR) coefficients were used to represent EEG signals of different sleep and wake stages. The proposed fast classification method was based on partial least squares regression (PLS), which was used to classify these features by finding an optimum beta using K-fold cross validation. The Physionet database was used to confirm accuracy and speed of the proposed classification system. This system could be used in real-time implementations because of its high classification rate, speed and capability to be implemented on hardware owing to be very comfortable. Finally, results of the PLS were compared with those of other classifiers such as k-nearest neighborhood (k-NN), linear discriminant classifier (LDC) and Bayes. We achieved 91% classification accuracy by selecting PLS as the classifier. These comparisons revealed that the proposed algorithm could recognize an emergency situation in less than 1s with high accuracy.

A new method for selection optimum k value in k-NN classification algorithm

In this paper a new algorithm to calculate optimum value of k for k-nearest neighborhood (k-NN) is proposed. Selection of k value is very important in k-NN classification algorithm. Our algorithm applied to sub-sampling and K-fold cross validation methods, separately. We applied our algorithm in different distribution of data set with different variances and means. We compared our algorithm with other classical k selection algorithms. The results show that the proposed algorithm achieved better performance than the classical algorithms.

Fast and high accuracy classification of sleep EEG using PLSR method

The aim of this study is to classify the status of sleep from electroencefelography (EEG) data recorded from seven different healthy individuals. The twenty two autoregressive (AR) model coefficent are computed and used as features. Three classification algorithms, namely k-NN, Bayes and PLSR methods are trained and tested. The results show that the PLSR algorithm yielded highest accuracy and short classification times. Furthermore, all utilizies just a single channel. Based on these results we propose that method can be used in clinical applications.

Observation of the effect of the increase in luminance level on emotional assessment in EEG records using Welch method

The aim of this study is to investigate the effect of the difference in level of luminance of visual stimuli on the emotional evaluation (negative, positive, neutral) from electroencephalogram (EEG) data recorded. EEG records of 31 healthy individuals were used in the study. These records were analyzed on the frequency domain using the Welch method and the features obtained after the analysis were evaluated statistically. According to the results of the evaluation, it was seen that there was a comparable difference between the reaction of individuals to the bright photographs with increased luminance level and the reaction of individuals to the original photographs, and it is revealed that this difference is spatially observed in which electrode region and in which frequency bands.

Investigation of luminance effect on the emotional assessment using Hjorth descriptors

The aim of this study is to investigate the impact of luminance level on the emotional evaluation from electroencephalography (EEG) records. EEG recordings obtained from 31 healty partipicants were used in this study. Featıres were obtained from EEG recordings using short time fourier transform (STFT) and Hjorth descriptors. According to the performance evaluation results (accuracy percentage), the value of the attributes of 70.97% of partipicants obtained from EEG recordings using the ORIGINAL negative visual stimule to be greater than the values of the attributes obtained from EEG recordings using the BRIGHT negative visual stimule were found. This observation was observed in the O2 electrode spatially region and delta frequency. This study is the beginning research on this subject and it is considered the results of this study will shed light on the issue.

Determining the effect of luminance on the EEG measurement with STFT method

The aim of this study is to observe the impact on the emotional evaluation of luminance, which is a perceptual feature of visual stimuli from electroencefelography (EEG) data recorded. EEG recordings obtained from 13 healthy individuals in the 25-45 age range were used in the study. Features obtained analysis of these records with Short Time Fourier Transform (STFT) are evaluated by statistically. The increase in the luminance level of the neutral visual stimuli as a result of statistical findings are reflected in the EEG activity, when given a bright neutral visual stimuli, comparable rates of exchange between the first 0-200 ms and 200-400ms on the electrical activity in the brain across the bright neutral stimuli to be higher than originally neutral stimulus were observed. This study has an initial research on the issue. In the following study, the difference in luminance level which regions of the brain, frequency and time, it will try to determine what caused the change.

Success of ensemble algorithms in classification of electrical impadence spectroscopy breast tissue records

In this study was performed by using records from breast tissue electrical impedance spectroscopy analysis. The aim of the study is to reveal the impact of ensemble algorithms on success of the classification performance in the classification of normal and pathological breast tissue classification. For this purpose have been used three different ensemble algorithms they are bagging, adaboost, random subspaces and three main basic classifiers, which are RF, YSA, DVM. The results obtained are supplemented with performance analysis and ensemble algorithms have been demonstrated to increase classification performance results. The results obtained by the combined use of adaboost ensemble algorithm with RF basic classifier demonstrate, that the success rate was higher than the others (%89.62).