SlideShare a Scribd company logo
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 836
DISENTANGLING BRAIN ACTIVITY FROM EEG DATA USING LOGISTIC
REGRESSION, XGBOOST, RNN AND CLASSIFICATION TREES.
Akshat Katiyar1
1UG Scholar, Dept. of CSE, PSG College of Technology, Tamil Nadu, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - EEG, is the method of choice to recordtheelectrical
activity generated by the brain via electrodes placed on the
scalp surface. Electrodes are mountedinelasticcapssimilarto
bathing caps, ensuring that the data can be collected from
identical scalp positions across all respondents. This project
aims at predicting a person’s learning capability using EEG
signals. The curiosity to learn about the changes that happen
inside our brain while doing certain tasks drives this research.
This research motivated us in developing a monitoringsystem
which uses Electroencephalogram (EEG) as a fundamental
physiological signal, to analyze and predict the learning
capability of a person. The primary focus of this study is to
identify the correlation between different data values
extracted from raw EEG signal and how they change while
doing different activities. This project disentangles brain
activity through EEG signal data of 10 college students and
predicts if a given student is confused or not while learning
new unknown topics. The project aims at predicting the
learning capabilities of an individual using EEG data from his
brain.
Key Words: Electroencephalogram, XGBoost, Logistic
Regression, Classification Trees
1. INTRODUCTION
An electroencephalogram (EEG)isa processusedtomonitor
electrical activity of the brain. An EEG tracks and records
brain wave patterns. Metal discs with thin wires are placed
on the scalp, and then send signals to a computer to record
the results. Normal electrical activity in the brain makes a
recognizable pattern.
1.1 EEG DATA EXTRACTION
To extract the brain waves signals there are multiple neural
waves recording devices in the market. For example,
Mindwave mobile produced andmarketed byNeuroSky.The
devices like these are tiny in size and can be put on and
removed easily and is connected to the software via
Bluetooth. In this case MyndPlayer Pro Ver 2.3 by Neurosky
was used as the software. The software canrecordtheactual
waveform of brain wave and the frequency of the various
components of the waveform. Thecomponentsareclassified
depending upon their frequency range. For an instance δ
waves (0.5 Hz to 3 Hz), θ waves (4 Hz to7Hz), midrange γ
waves (41 Hz to 50 Hz). In our case, we have 10 students
who watch multiple 2-minute videos. We collect our
experimenting data when the focus and the attention of the
student is maximum. We remove the first and last 30
seconds. The different variations of the brain nerves during
this activity is captured by the above-mentioned software.
Figure 1.1 Variations of Brain Nerves
1.2 PREPROCESSING OF EEG SIGNALS
To interpret the signals that are fed into the computers the
signals has to be processed. The EEG signals consists of
several nonlinear distinct waveforms known as band
components. The band components are complex and are
categorized by the frequency range they fall into. The
different band components are extracted from the raw EEG
signals (Butterworth bandpass filters).
The following are the five primary bands of EEG signals:
 Delta (0.2–4Hz)
 Theta (4– 8 Hz)
 Alpha (8–13 Hz)
 Beta (13–30 Hz)
 Gamma (30– 55 Hz)
The small yet complex varying frequency structure found in
scalp-recorded EEG waveforms contains detailed
neuroelectric information. To analyze and isolate such
waveforms or rhythms wavelet analysis is used. Since
wavelet coefficients allow the precise noise filtering by
attenuating the coefficients associated primarily with noise
before reconstructing the signal with wavelet synthesis it
has been used in the removal or separation of noisefromthe
raw EEG waveforms.
There are two types of wavelet transforms: Continuous
wavelet transform (CWT) and Discrete wavelet transforms
(DWT). We in our experiment apply Discrete wavelet
transform because it allows the analysis of signals by
applying only discrete values of shift and scaling to form the
discrete wavelets. The other advantage of applying DWT is
that if suppose the original signal is sampled with a suitable
set of shifting and scaling values, using the inverse of DWT
the entire continuous signal can be reconstructed. The
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 837
general Discrete wavelet representation is shown below:
Where,
 integer’s m and n control the wavelet shifting and
scaling, respectively,
 a0 is a specified fixed dilation step parameter set at
a value greater than 1,
 b0 is the location parameter which must be greater
than zero.
Figure 1.2 EEG Decomposition and Extraction
The Figure 1.2 shows the EEG decomposition and extraction
using the wavelet transformation. As shown in the figure
each of the extracted band components are sent through the
“Windowing step”. In this step the interesting and boring
portions of the bandcomponents basedonthetimestampsof
the original EEG are extractedandsentthrougha windowing
mechanism. In this mechanism, each bandcomponentsignal
is partitioned into tiny windows.Afterthisstep eachwindow
is decomposed using DWT. The decomposition process in
wavelet transform can be performed iteratively into several
levels. The values of the coefficients from the window
interval is considered as the extracted features of EEG
signals.
1.3 BRAIN WAVES CLASSIFICATION
Brain patterns form wave shapes that are commonly
sinusoidal. Usually, they are measured from peak to peak
and normally range from 0.5 to 100 µV in amplitude, which
is about 100 times lower than ECG signals. By means of
Fourier transform powerspectrum fromtherawEEGsignal
is derived. In power spectrum contribution of sine waves
with different frequenciesarevisible.Althoughthespectrum
is continuous, ranging from 0 Hz up to one half of sampling
frequency, the brain state of the individual maymakecertain
frequencies more dominant.
Brain waves have been categorized into four basic groups:
Figure 1.3 Different Categories of Brain Waves
2. SYSTEM DESIGN
The schematic diagram illustrates the overall method ofthis
project. As shown EEG signals areacquiredfromthesubjects
during the experiment.
In the project we examine 10 students who watch
educational videos of two kinds (Confusing one and the one
which is from an easy topic). The duration of the video is 2
minutes and we take voltage difference of the waveforms
emitted by the brains of the students.
Thus, the brain wave pattern is monitored using electrodes
and sensors. After the activity of watching the video,
students rate their level of confusion, either 1 or 0.’0’ means
the topic is not confusing and the value ‘1’ means the topic
was new or surprising to them.
The entire video of 2 minutes is cut into 1 min (by removing
the first 30s and last 30s). And the response fromstudentsis
taken for each 5s session. Thus, getting 12 modes for one
video for one user.
We have two .csv files as our datasets in this project. The
first one gives us the demographic information about the
students and it is not much of use to us, since we perform
our operations and apply our algorithms on the second file,
which gives us the EEG signal information.
We initially perform our description analysis on this dataset
by viewing the number of rows and columns present in it
and also checking for missing values if any. There are
actually 15 columns out of which 4 of them(studentid,video
id, user defined label, pre - defined label) are converted into
integer for us to group and analyze the dataset easily.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 838
The 4 described columns will also not affect the
understanding of the particular subject for the student. So,
they are included in the process of feature extraction. Since,
all the features are of numerical data type (floating point
numbers), we apply Pearson’s coefficient as our tool. After
applying, the Pearson’s coefficient we find the relationship
between the features by plotting them in different plots
(Cross tables, bar graphs and heatmaps). Later we apply
Logistic Regression, RNN, Boost and Decision Tree on the
same.
Figure 2 Comparison of Bar Graphs
We can infer from the above graph (Figure 3.a) that, video 9
was the easiest video since, only 2 students have told the
video was surprising to them. In the same way, subject 1 has
more learning capability than others since, he found only 4
videos were surprising.
2.1 DATA PREPROCESSING AND EXTRACTION
The EEG signal is comprised of a complex and nonlinear
combination of several distinct waveforms which are also
called band components. Each of the band components is
categorized by the frequency range that they exist in. The
state of consciousness of the individuals may make one
frequency range more pronounced than others. As shown in
Figure above, the different band components are extracted
from the raw EEG signal using Butterworth bandpass filters.
Five primary bands of the EEG signal are extracted, namely,
Delta (0.2–4Hz), Theta (4– 8Hz), Alpha (8–13Hz), Beta (13–
30Hz), and Gamma (30– 55Hz).
This Section tells about how key features are extracted from
raw EEG signal. This Sample EEG data is of two subjects. It’s
given in a .EDF file which is basically Raw signal. The Output
of such file is given below
Figure 2.1 Raw EEG Signals
The above .EDF file representation (Figure 2.1) has 7 signals
in which 2 are EEG channels.
2.2 FEATURE SELECTION
Feature selection is a process whereyouautomaticallyselect
those features in your data that contribute most to the
prediction variable or output in which you are interested.
Types of feature Selection performed
 Univariate Selection (operates only on positivevariables)
 Recursive Feature Selection
 Pearson Correlation
 Feature Importance
2.2.1 UNIVARIATE SELECTION
Univariate Selection is a set of Statistical tests to selectthose
features that have the strongest relationship withtheoutput
variable. Univariate Selection Works only on Positive
Variable as we have a feature Raw which has negative
values. The remaining 10 features are used for the
classification. After Selecting the 10 features we use
SelectKBest and chi2 to find the best features among the
selected 10. Later the most Dominant Features are used for
Classification using Decision Tree, Logistic Regression,
Extreme Gradient Boosting.
2.2.2 RECURSIVE FEATURE SELECTION
Recursive feature elimination is a feature selection method
that fits a model and removes the weakest feature (or set of
features) until the specified number of features is reached.
This method works by recursively selecting attributes and
building a model on those attributes that remain. After
selecting the 10 features we use RFE to findthe bestfeatures
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 839
among the selected 10.
2.2.3 PEARSON CORRELATION
When a data set has too many variables. Most of the
variables are correlated. So, running a model on whole data
returns poor accuracy. Finding the most positively
correlated variables is mostly the best choice. Because they
affect the output the most. Correlation is the mutual
relationship between two features.
After Selecting the 11 features we use correlation (Pearson)
to find the best features among the selected 11. Later the
most Dominant Features are used for Classification using
Decision Tree, Logistic Regression, Extreme Gradient
Boosting.
Figure 2.2.3 Pearson Correlation
Using Pearson Correlation, the heat map looks like this
(Figure 2.2.3) it shows Delta, Alpha, Gamma, Beta, Theta are
strongly correlated and are most dominant.
2.2.4 FEATURE IMPORTANCE
A common approach to eliminating features is to describe
their relative importance to a model, then eliminate weak
features or combinations of features andre-evaluatetoseeif
the model fairs better during cross-validation this is feature
importance. This method gives the feature importance of
each feature of the dataset by using the feature importance
property of the model.Featureimportancegivesthescorefor
each feature of your data, the higher the score more
important or relevant is the feature towards the output
variable. After Selecting the 11 features we use
ExtraTreesClassifier to find the best features among the
selected 11. Later the most Dominant Features are used for
Classification using Decision Tree, Logistic Regression,
Extreme Gradient Boosting.
2.2.4 RAW SLEEP EEG DATASET
The Sleep Physio net dataset is annotated using 8 labels
<physionet_labels> Wake (W),Stage1,Stage2,Stage3, Stage
4 corresponding to the range from light sleep to deep sleep,
REM sleep (R) where REM is the abbreviation for Rapid Eye
Movement sleep, movement (M), and Stage (?) for any none
scored segment.
Figure 2.2.4 Sleep Stages
We will work only with 5 stages: Wake (W), Stage 1, Stage 2,
Stage 3/4, and REM sleep (R). Here Stage 1, Stage 2, Stage
3/4 they are nothing just the sleep concentration where
Stage 1 has less sleep concentration than Stage 2.
This is the Sleep data collected from a person (16 hrs. data).
It has following values.
Where FpZ (Intersection between Fz and Fp1 and Fp2 in the
figure), Fz, Cz, Oz is just electrode labelling
2.2.5 DEPENDENT FEATURES(GRAPH)
Here each abbreviation means.
A-Attention, M-Meditation, R-Raw, D-Delta, A1-Alpha1, A2-
Alpha2, B1-Beta1, B2-Beta2, G1-Gamma1, G2-Gamma2
The set of features dependent on each other is as
follows:
MAIN FEATURE DEPENDENT FEATURE
R {A1, A2, B1, B2, G1, G2}
{A1, A2, G1, D, M} A
{G2, A, D} M
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 840
From the above graph following inferences are made. A
(attention) is most dependent among all, R (Raw) is the
origin of all-important feature like alpha, gamma etc. Hence,
we conclude Attention is the most dominant feature in our
study.
3. RESULT
After the code was completed various parameters were
tuned for the best results. Adding all the features provided
more accuracy than with using selected features. Classifiers
were also changed to see which gives the best result. Here is
the list of Classifier tried and the corresponding accuracies
obtained with their corresponding run-time.Classificationis
one of the methods used to identify the certain pattern from
feature extraction such as amplitude of EEG data, power
spectral density (PSD), and Band power (BP). The
performances of identify the pattern are depend on both
features and classification algorithm. There are many
classification algorithms used in the EEG analysis. Here
following classifiers are used.
3.1 COMPARISON
 Better Dataset with high amount of training data.
 Independent features at least one so that all
features are not needed for classification.
 Adding two new columns specifying the level of
difficulty to make predictions more accurate.
The Figure 3.1(a) shows 3 algorithms along with their run-
time, clearly taking all features gives the best accuracy but it
does take a lot of time. Also, you can see XGBoost takes more
time than the other two because it keeps searching for local
minima and then finds a common global minimum. Hence it
takes more time than other two classifiers.
Figure 3.1(a) Comparison of Classifiers
Figure 3.1(b) RNN Accuracy
The Figure 3.1(b) shows the accuracy of RNN. As seen after
taking hidden layers greater than 10 the model goes to a
saturation state and the accuracy remains same.
3. CONCLUSIONS
Predicting the learning capabilities of an individual is not an
easy task as it has lot of complex tasks involved. It can be
seen that reasonable amount of accuracy achieved when all
the features are selected. Amongst the various machine
learning models investigated the classification using
Extreme gradient boosting seems toperformbest(63.6%)
on the combined feature set. However logistic regression
using all features and recursive feature selection gives the
closest accuracy to the best(61.3&, 61.2%).Also,byapplying
various feature selection algorithms on the dataset wecome
to know that we get the maximum accuracy of prediction
when all the features are selected. This not only
demonstrates that our dataset is small but also depicts that
for making the prediction we can’t depend only on our four
important features(theta,alpha,beta,gamma).Although this
work is limited to EEG data there is a room for improvement
using clinical data and genetic data.
Here are some of the future work planned to improve the
system’s classification and prediction performance.
 A larger data set is needed to further validate this
experiment. A larger data set is expected to provide
a more robust classifier model.
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072
© 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 841
 Having a more diverse base of features usually
provides insight into some connate characteristics
of the signal which might not be openly evident. 

 Feature pruning and other classification methods
need to be tried for increasing the accuracy. 

Hence this project gives a little insight of how brain react to
certain changes and how those changes can be used to
predict certain tasks.
REFERENCES
[1] T. Blakely, K.J. Miller, R. P N Rao, Mark D.Holmes,andJ.G.
Ojemann, “Localization and classification of phonemes
using high spatial resolution electrocortico
graphy(ECoG) grids,” in Engineering in Medicine and
Biology Society, 2008. EMBS 2008. 30th Annual
International Conference of the IEEE, Aug 2008,
pp.4964–4967.
[2] Spencer Kellis, Kai Miller, Kyle Thomson, Richard
Brown, Paul House, and Bradley Greger, “Decoding
spoken words using local field potentials recordedfrom
thecortical surface,” Journal of Neural Engineering, vol.
7, no. 5, pp.1–10,2010.
[3] Jess Bartels, Dinal Andreasen, Princewill Ehirim,
HuiMao, Steven Seibert, E. Joe Wright, and Philip
Kennedy,“Neurotrophic electrode: Method of assembly
and implantation into human motor speech cortex,”
Journal of Neuro science Methods, vol. 174, no. 2,
pp.168–176,2008.
[4] https://www.kdnuggets.com/2017/10/xgboost-top-
machine-learning-method-kaggle-explaind.html
[5] https://physionet.org/physiobank/database/sleep-
edfx/sleep-cassette

More Related Content

What's hot

E44082429
E44082429E44082429
E44082429
IJERA Editor
 
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORKEEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
IJCI JOURNAL
 
Introduction to Artificial Neural Networks
Introduction to Artificial Neural NetworksIntroduction to Artificial Neural Networks
Introduction to Artificial Neural Networks
Adri Jovin
 
IRJET- Review on Depression Prediction using Different Methods
IRJET- Review on Depression Prediction using Different MethodsIRJET- Review on Depression Prediction using Different Methods
IRJET- Review on Depression Prediction using Different Methods
IRJET Journal
 
Using Brain Waves as New Biometric Feature for Authenticating a Computer User...
Using Brain Waves as New Biometric Feature for Authenticatinga Computer User...Using Brain Waves as New Biometric Feature for Authenticatinga Computer User...
Using Brain Waves as New Biometric Feature for Authenticating a Computer User...
Buthainah Hamdy
 
IRJET- Mental Workload Assessment using Rnn
IRJET- Mental Workload Assessment using RnnIRJET- Mental Workload Assessment using Rnn
IRJET- Mental Workload Assessment using Rnn
IRJET Journal
 
PID1563185
PID1563185PID1563185
PID1563185
Krishna Potnuru
 
IRJET- Depression Prediction System using Different Methods
IRJET- Depression Prediction System using Different MethodsIRJET- Depression Prediction System using Different Methods
IRJET- Depression Prediction System using Different Methods
IRJET Journal
 
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
CSCJournals
 
I365358
I365358I365358
I365358
IJERA Editor
 
06201565
0620156506201565
06201565
rthara
 
11.simulation of unified power quality conditioner for power quality improvem...
11.simulation of unified power quality conditioner for power quality improvem...11.simulation of unified power quality conditioner for power quality improvem...
11.simulation of unified power quality conditioner for power quality improvem...
Alexander Decker
 
Artificial neural networks
Artificial neural networks Artificial neural networks
Artificial neural networks
Institute of Technology Telkom
 
Unit+i
Unit+iUnit+i
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Nurhasanah Shafei
 
238 243
238 243238 243
Eyeblink artefact removal from eeg using independent
Eyeblink artefact removal from eeg using independentEyeblink artefact removal from eeg using independent
Eyeblink artefact removal from eeg using independent
eSAT Publishing House
 
Neural Network
Neural NetworkNeural Network
Neural Network
Ashish Kumar
 
ANN-lecture9
ANN-lecture9ANN-lecture9
ANN-lecture9
Laila Fatehy
 
Advanced applications of artificial intelligence and neural networks
Advanced applications of artificial intelligence and neural networksAdvanced applications of artificial intelligence and neural networks
Advanced applications of artificial intelligence and neural networks
Praveen Kumar
 

What's hot (20)

E44082429
E44082429E44082429
E44082429
 
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORKEEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
EEG SIGNAL IDENTIFICATION USING SINGLE-LAYER NEURAL NETWORK
 
Introduction to Artificial Neural Networks
Introduction to Artificial Neural NetworksIntroduction to Artificial Neural Networks
Introduction to Artificial Neural Networks
 
IRJET- Review on Depression Prediction using Different Methods
IRJET- Review on Depression Prediction using Different MethodsIRJET- Review on Depression Prediction using Different Methods
IRJET- Review on Depression Prediction using Different Methods
 
Using Brain Waves as New Biometric Feature for Authenticating a Computer User...
Using Brain Waves as New Biometric Feature for Authenticatinga Computer User...Using Brain Waves as New Biometric Feature for Authenticatinga Computer User...
Using Brain Waves as New Biometric Feature for Authenticating a Computer User...
 
IRJET- Mental Workload Assessment using Rnn
IRJET- Mental Workload Assessment using RnnIRJET- Mental Workload Assessment using Rnn
IRJET- Mental Workload Assessment using Rnn
 
PID1563185
PID1563185PID1563185
PID1563185
 
IRJET- Depression Prediction System using Different Methods
IRJET- Depression Prediction System using Different MethodsIRJET- Depression Prediction System using Different Methods
IRJET- Depression Prediction System using Different Methods
 
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
A New Approach to Denoising EEG Signals - Merger of Translation Invariant Wav...
 
I365358
I365358I365358
I365358
 
06201565
0620156506201565
06201565
 
11.simulation of unified power quality conditioner for power quality improvem...
11.simulation of unified power quality conditioner for power quality improvem...11.simulation of unified power quality conditioner for power quality improvem...
11.simulation of unified power quality conditioner for power quality improvem...
 
Artificial neural networks
Artificial neural networks Artificial neural networks
Artificial neural networks
 
Unit+i
Unit+iUnit+i
Unit+i
 
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.Dsp lab report- Analysis and classification of EMG signal using MATLAB.
Dsp lab report- Analysis and classification of EMG signal using MATLAB.
 
238 243
238 243238 243
238 243
 
Eyeblink artefact removal from eeg using independent
Eyeblink artefact removal from eeg using independentEyeblink artefact removal from eeg using independent
Eyeblink artefact removal from eeg using independent
 
Neural Network
Neural NetworkNeural Network
Neural Network
 
ANN-lecture9
ANN-lecture9ANN-lecture9
ANN-lecture9
 
Advanced applications of artificial intelligence and neural networks
Advanced applications of artificial intelligence and neural networksAdvanced applications of artificial intelligence and neural networks
Advanced applications of artificial intelligence and neural networks
 

Similar to IRJET- Disentangling Brain Activity from EEG Data using Logistic Regression, XGBOOST, RNN and Classification Trees.

Epileptic Seizure Detection using An EEG Sensor
Epileptic Seizure Detection using An EEG SensorEpileptic Seizure Detection using An EEG Sensor
Epileptic Seizure Detection using An EEG Sensor
IRJET Journal
 
IRJET-A Survey on Effect of Meditation on Attention Level Using EEG
IRJET-A Survey on Effect of Meditation on Attention Level Using EEGIRJET-A Survey on Effect of Meditation on Attention Level Using EEG
IRJET-A Survey on Effect of Meditation on Attention Level Using EEG
IRJET Journal
 
The International Journal of Engineering and Science (IJES)
The International Journal of Engineering and Science (IJES)The International Journal of Engineering and Science (IJES)
The International Journal of Engineering and Science (IJES)
theijes
 
IRJET- Analysis of Electroencephalogram (EEG) Signals
IRJET- Analysis of Electroencephalogram (EEG) SignalsIRJET- Analysis of Electroencephalogram (EEG) Signals
IRJET- Analysis of Electroencephalogram (EEG) Signals
IRJET Journal
 
Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...
TELKOMNIKA JOURNAL
 
Denoising Techniques for EEG Signals: A Review
Denoising Techniques for EEG Signals: A ReviewDenoising Techniques for EEG Signals: A Review
Denoising Techniques for EEG Signals: A Review
IRJET Journal
 
STRESS ANALYSIS
STRESS ANALYSISSTRESS ANALYSIS
STRESS ANALYSIS
shiva karthik reddy koyya
 
Wavelet Based Feature Extraction Scheme Of Eeg Waveform
Wavelet Based Feature Extraction Scheme Of Eeg WaveformWavelet Based Feature Extraction Scheme Of Eeg Waveform
Wavelet Based Feature Extraction Scheme Of Eeg Waveform
shan pri
 
Biomedical Signals Classification With Transformer Based Model.pptx
Biomedical Signals Classification With Transformer Based Model.pptxBiomedical Signals Classification With Transformer Based Model.pptx
Biomedical Signals Classification With Transformer Based Model.pptx
Sandeep Kumar
 
Transfer learning for epilepsy detection using spectrogram images
Transfer learning for epilepsy detection using spectrogram imagesTransfer learning for epilepsy detection using spectrogram images
Transfer learning for epilepsy detection using spectrogram images
IAESIJAI
 
Prediction Model for Emotion Recognition Using EEG
Prediction Model for Emotion Recognition Using EEGPrediction Model for Emotion Recognition Using EEG
Prediction Model for Emotion Recognition Using EEG
IRJET Journal
 
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
IJCSEA Journal
 
40120130405007
4012013040500740120130405007
40120130405007
IAEME Publication
 
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy ComparisonAnalysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
ijsrd.com
 
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
IAESIJEECS
 
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
Editor IJCATR
 
EEG S IGNAL Q UANTIFICATION B ASED ON M ODUL L EVELS
EEG S IGNAL  Q UANTIFICATION  B ASED ON M ODUL  L EVELS EEG S IGNAL  Q UANTIFICATION  B ASED ON M ODUL  L EVELS
EEG S IGNAL Q UANTIFICATION B ASED ON M ODUL L EVELS
sipij
 
IRJET-Electromyogram Signals for Multiuser Interface- A Review
IRJET-Electromyogram Signals for Multiuser Interface- A ReviewIRJET-Electromyogram Signals for Multiuser Interface- A Review
IRJET-Electromyogram Signals for Multiuser Interface- A Review
IRJET Journal
 
EEG Signal Classification using Deep Neural Network
EEG Signal Classification using Deep Neural NetworkEEG Signal Classification using Deep Neural Network
EEG Signal Classification using Deep Neural Network
IRJET Journal
 
IRJET-Estimation of Meditation Effect on Attention Level using EEG
IRJET-Estimation of Meditation Effect on Attention Level using EEGIRJET-Estimation of Meditation Effect on Attention Level using EEG
IRJET-Estimation of Meditation Effect on Attention Level using EEG
IRJET Journal
 

Similar to IRJET- Disentangling Brain Activity from EEG Data using Logistic Regression, XGBOOST, RNN and Classification Trees. (20)

Epileptic Seizure Detection using An EEG Sensor
Epileptic Seizure Detection using An EEG SensorEpileptic Seizure Detection using An EEG Sensor
Epileptic Seizure Detection using An EEG Sensor
 
IRJET-A Survey on Effect of Meditation on Attention Level Using EEG
IRJET-A Survey on Effect of Meditation on Attention Level Using EEGIRJET-A Survey on Effect of Meditation on Attention Level Using EEG
IRJET-A Survey on Effect of Meditation on Attention Level Using EEG
 
The International Journal of Engineering and Science (IJES)
The International Journal of Engineering and Science (IJES)The International Journal of Engineering and Science (IJES)
The International Journal of Engineering and Science (IJES)
 
IRJET- Analysis of Electroencephalogram (EEG) Signals
IRJET- Analysis of Electroencephalogram (EEG) SignalsIRJET- Analysis of Electroencephalogram (EEG) Signals
IRJET- Analysis of Electroencephalogram (EEG) Signals
 
Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...Brain-computer interface of focus and motor imagery using wavelet and recurre...
Brain-computer interface of focus and motor imagery using wavelet and recurre...
 
Denoising Techniques for EEG Signals: A Review
Denoising Techniques for EEG Signals: A ReviewDenoising Techniques for EEG Signals: A Review
Denoising Techniques for EEG Signals: A Review
 
STRESS ANALYSIS
STRESS ANALYSISSTRESS ANALYSIS
STRESS ANALYSIS
 
Wavelet Based Feature Extraction Scheme Of Eeg Waveform
Wavelet Based Feature Extraction Scheme Of Eeg WaveformWavelet Based Feature Extraction Scheme Of Eeg Waveform
Wavelet Based Feature Extraction Scheme Of Eeg Waveform
 
Biomedical Signals Classification With Transformer Based Model.pptx
Biomedical Signals Classification With Transformer Based Model.pptxBiomedical Signals Classification With Transformer Based Model.pptx
Biomedical Signals Classification With Transformer Based Model.pptx
 
Transfer learning for epilepsy detection using spectrogram images
Transfer learning for epilepsy detection using spectrogram imagesTransfer learning for epilepsy detection using spectrogram images
Transfer learning for epilepsy detection using spectrogram images
 
Prediction Model for Emotion Recognition Using EEG
Prediction Model for Emotion Recognition Using EEGPrediction Model for Emotion Recognition Using EEG
Prediction Model for Emotion Recognition Using EEG
 
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
ENERGY COMPUTATION FOR BCI USING DCT AND MOVING AVERAGE WINDOW FOR NOISE SMOO...
 
40120130405007
4012013040500740120130405007
40120130405007
 
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy ComparisonAnalysis of EEG data Using ICA and Algorithm Development for Energy Comparison
Analysis of EEG data Using ICA and Algorithm Development for Energy Comparison
 
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
5. detection and separation of eeg artifacts using wavelet transform nov 11, ...
 
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
Feature Extraction Techniques and Classification Algorithms for EEG Signals t...
 
EEG S IGNAL Q UANTIFICATION B ASED ON M ODUL L EVELS
EEG S IGNAL  Q UANTIFICATION  B ASED ON M ODUL  L EVELS EEG S IGNAL  Q UANTIFICATION  B ASED ON M ODUL  L EVELS
EEG S IGNAL Q UANTIFICATION B ASED ON M ODUL L EVELS
 
IRJET-Electromyogram Signals for Multiuser Interface- A Review
IRJET-Electromyogram Signals for Multiuser Interface- A ReviewIRJET-Electromyogram Signals for Multiuser Interface- A Review
IRJET-Electromyogram Signals for Multiuser Interface- A Review
 
EEG Signal Classification using Deep Neural Network
EEG Signal Classification using Deep Neural NetworkEEG Signal Classification using Deep Neural Network
EEG Signal Classification using Deep Neural Network
 
IRJET-Estimation of Meditation Effect on Attention Level using EEG
IRJET-Estimation of Meditation Effect on Attention Level using EEGIRJET-Estimation of Meditation Effect on Attention Level using EEG
IRJET-Estimation of Meditation Effect on Attention Level using EEG
 

More from IRJET Journal

TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
IRJET Journal
 
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURESTUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
IRJET Journal
 
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
IRJET Journal
 
Effect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil CharacteristicsEffect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil Characteristics
IRJET Journal
 
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
IRJET Journal
 
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
IRJET Journal
 
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
IRJET Journal
 
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
IRJET Journal
 
A REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADASA REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADAS
IRJET Journal
 
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
IRJET Journal
 
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD ProP.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
IRJET Journal
 
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
IRJET Journal
 
Survey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare SystemSurvey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare System
IRJET Journal
 
Review on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridgesReview on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridges
IRJET Journal
 
React based fullstack edtech web application
React based fullstack edtech web applicationReact based fullstack edtech web application
React based fullstack edtech web application
IRJET Journal
 
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
IRJET Journal
 
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
IRJET Journal
 
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
IRJET Journal
 
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic DesignMultistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
IRJET Journal
 
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
IRJET Journal
 

More from IRJET Journal (20)

TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
TUNNELING IN HIMALAYAS WITH NATM METHOD: A SPECIAL REFERENCES TO SUNGAL TUNNE...
 
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURESTUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
STUDY THE EFFECT OF RESPONSE REDUCTION FACTOR ON RC FRAMED STRUCTURE
 
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
A COMPARATIVE ANALYSIS OF RCC ELEMENT OF SLAB WITH STARK STEEL (HYSD STEEL) A...
 
Effect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil CharacteristicsEffect of Camber and Angles of Attack on Airfoil Characteristics
Effect of Camber and Angles of Attack on Airfoil Characteristics
 
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
A Review on the Progress and Challenges of Aluminum-Based Metal Matrix Compos...
 
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
Dynamic Urban Transit Optimization: A Graph Neural Network Approach for Real-...
 
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
Structural Analysis and Design of Multi-Storey Symmetric and Asymmetric Shape...
 
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
A Review of “Seismic Response of RC Structures Having Plan and Vertical Irreg...
 
A REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADASA REVIEW ON MACHINE LEARNING IN ADAS
A REVIEW ON MACHINE LEARNING IN ADAS
 
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
Long Term Trend Analysis of Precipitation and Temperature for Asosa district,...
 
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD ProP.E.B. Framed Structure Design and Analysis Using STAAD Pro
P.E.B. Framed Structure Design and Analysis Using STAAD Pro
 
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
A Review on Innovative Fiber Integration for Enhanced Reinforcement of Concre...
 
Survey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare SystemSurvey Paper on Cloud-Based Secured Healthcare System
Survey Paper on Cloud-Based Secured Healthcare System
 
Review on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridgesReview on studies and research on widening of existing concrete bridges
Review on studies and research on widening of existing concrete bridges
 
React based fullstack edtech web application
React based fullstack edtech web applicationReact based fullstack edtech web application
React based fullstack edtech web application
 
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
A Comprehensive Review of Integrating IoT and Blockchain Technologies in the ...
 
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
A REVIEW ON THE PERFORMANCE OF COCONUT FIBRE REINFORCED CONCRETE.
 
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
Optimizing Business Management Process Workflows: The Dynamic Influence of Mi...
 
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic DesignMultistoried and Multi Bay Steel Building Frame by using Seismic Design
Multistoried and Multi Bay Steel Building Frame by using Seismic Design
 
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
Cost Optimization of Construction Using Plastic Waste as a Sustainable Constr...
 

Recently uploaded

CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1
PKavitha10
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
shadow0702a
 
People as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimalaPeople as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimala
riddhimaagrawal986
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
IJECEIAES
 
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by AnantLLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
Anant Corporation
 
AI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptxAI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptx
architagupta876
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
co23btech11018
 
Properties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptxProperties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptx
MDSABBIROJJAMANPAYEL
 
integral complex analysis chapter 06 .pdf
integral complex analysis chapter 06 .pdfintegral complex analysis chapter 06 .pdf
integral complex analysis chapter 06 .pdf
gaafergoudaay7aga
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
VICTOR MAESTRE RAMIREZ
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
171ticu
 
Data Control Language.pptx Data Control Language.pptx
Data Control Language.pptx Data Control Language.pptxData Control Language.pptx Data Control Language.pptx
Data Control Language.pptx Data Control Language.pptx
ramrag33
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
KrishnaveniKrishnara1
 
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
IJECEIAES
 
Engineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdfEngineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdf
abbyasa1014
 
ITSM Integration with MuleSoft.pptx
ITSM  Integration with MuleSoft.pptxITSM  Integration with MuleSoft.pptx
ITSM Integration with MuleSoft.pptx
VANDANAMOHANGOUDA
 
Null Bangalore | Pentesters Approach to AWS IAM
Null Bangalore | Pentesters Approach to AWS IAMNull Bangalore | Pentesters Approach to AWS IAM
Null Bangalore | Pentesters Approach to AWS IAM
Divyanshu
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
Yasser Mahgoub
 
cnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classicationcnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classication
SakkaravarthiShanmug
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
Gino153088
 

Recently uploaded (20)

CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1
 
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
Use PyCharm for remote debugging of WSL on a Windo cf5c162d672e4e58b4dde5d797...
 
People as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimalaPeople as resource Grade IX.pdf minimala
People as resource Grade IX.pdf minimala
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
 
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by AnantLLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
 
AI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptxAI assisted telemedicine KIOSK for Rural India.pptx
AI assisted telemedicine KIOSK for Rural India.pptx
 
Computational Engineering IITH Presentation
Computational Engineering IITH PresentationComputational Engineering IITH Presentation
Computational Engineering IITH Presentation
 
Properties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptxProperties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptx
 
integral complex analysis chapter 06 .pdf
integral complex analysis chapter 06 .pdfintegral complex analysis chapter 06 .pdf
integral complex analysis chapter 06 .pdf
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
 
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样官方认证美国密歇根州立大学毕业证学位证书原版一模一样
官方认证美国密歇根州立大学毕业证学位证书原版一模一样
 
Data Control Language.pptx Data Control Language.pptx
Data Control Language.pptx Data Control Language.pptxData Control Language.pptx Data Control Language.pptx
Data Control Language.pptx Data Control Language.pptx
 
22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt22CYT12-Unit-V-E Waste and its Management.ppt
22CYT12-Unit-V-E Waste and its Management.ppt
 
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
Redefining brain tumor segmentation: a cutting-edge convolutional neural netw...
 
Engineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdfEngineering Drawings Lecture Detail Drawings 2014.pdf
Engineering Drawings Lecture Detail Drawings 2014.pdf
 
ITSM Integration with MuleSoft.pptx
ITSM  Integration with MuleSoft.pptxITSM  Integration with MuleSoft.pptx
ITSM Integration with MuleSoft.pptx
 
Null Bangalore | Pentesters Approach to AWS IAM
Null Bangalore | Pentesters Approach to AWS IAMNull Bangalore | Pentesters Approach to AWS IAM
Null Bangalore | Pentesters Approach to AWS IAM
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
 
cnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classicationcnn.pptx Convolutional neural network used for image classication
cnn.pptx Convolutional neural network used for image classication
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
 

IRJET- Disentangling Brain Activity from EEG Data using Logistic Regression, XGBOOST, RNN and Classification Trees.

  • 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 836 DISENTANGLING BRAIN ACTIVITY FROM EEG DATA USING LOGISTIC REGRESSION, XGBOOST, RNN AND CLASSIFICATION TREES. Akshat Katiyar1 1UG Scholar, Dept. of CSE, PSG College of Technology, Tamil Nadu, India ---------------------------------------------------------------------***---------------------------------------------------------------------- Abstract - EEG, is the method of choice to recordtheelectrical activity generated by the brain via electrodes placed on the scalp surface. Electrodes are mountedinelasticcapssimilarto bathing caps, ensuring that the data can be collected from identical scalp positions across all respondents. This project aims at predicting a person’s learning capability using EEG signals. The curiosity to learn about the changes that happen inside our brain while doing certain tasks drives this research. This research motivated us in developing a monitoringsystem which uses Electroencephalogram (EEG) as a fundamental physiological signal, to analyze and predict the learning capability of a person. The primary focus of this study is to identify the correlation between different data values extracted from raw EEG signal and how they change while doing different activities. This project disentangles brain activity through EEG signal data of 10 college students and predicts if a given student is confused or not while learning new unknown topics. The project aims at predicting the learning capabilities of an individual using EEG data from his brain. Key Words: Electroencephalogram, XGBoost, Logistic Regression, Classification Trees 1. INTRODUCTION An electroencephalogram (EEG)isa processusedtomonitor electrical activity of the brain. An EEG tracks and records brain wave patterns. Metal discs with thin wires are placed on the scalp, and then send signals to a computer to record the results. Normal electrical activity in the brain makes a recognizable pattern. 1.1 EEG DATA EXTRACTION To extract the brain waves signals there are multiple neural waves recording devices in the market. For example, Mindwave mobile produced andmarketed byNeuroSky.The devices like these are tiny in size and can be put on and removed easily and is connected to the software via Bluetooth. In this case MyndPlayer Pro Ver 2.3 by Neurosky was used as the software. The software canrecordtheactual waveform of brain wave and the frequency of the various components of the waveform. Thecomponentsareclassified depending upon their frequency range. For an instance δ waves (0.5 Hz to 3 Hz), θ waves (4 Hz to7Hz), midrange γ waves (41 Hz to 50 Hz). In our case, we have 10 students who watch multiple 2-minute videos. We collect our experimenting data when the focus and the attention of the student is maximum. We remove the first and last 30 seconds. The different variations of the brain nerves during this activity is captured by the above-mentioned software. Figure 1.1 Variations of Brain Nerves 1.2 PREPROCESSING OF EEG SIGNALS To interpret the signals that are fed into the computers the signals has to be processed. The EEG signals consists of several nonlinear distinct waveforms known as band components. The band components are complex and are categorized by the frequency range they fall into. The different band components are extracted from the raw EEG signals (Butterworth bandpass filters). The following are the five primary bands of EEG signals:  Delta (0.2–4Hz)  Theta (4– 8 Hz)  Alpha (8–13 Hz)  Beta (13–30 Hz)  Gamma (30– 55 Hz) The small yet complex varying frequency structure found in scalp-recorded EEG waveforms contains detailed neuroelectric information. To analyze and isolate such waveforms or rhythms wavelet analysis is used. Since wavelet coefficients allow the precise noise filtering by attenuating the coefficients associated primarily with noise before reconstructing the signal with wavelet synthesis it has been used in the removal or separation of noisefromthe raw EEG waveforms. There are two types of wavelet transforms: Continuous wavelet transform (CWT) and Discrete wavelet transforms (DWT). We in our experiment apply Discrete wavelet transform because it allows the analysis of signals by applying only discrete values of shift and scaling to form the discrete wavelets. The other advantage of applying DWT is that if suppose the original signal is sampled with a suitable set of shifting and scaling values, using the inverse of DWT the entire continuous signal can be reconstructed. The
  • 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 837 general Discrete wavelet representation is shown below: Where,  integer’s m and n control the wavelet shifting and scaling, respectively,  a0 is a specified fixed dilation step parameter set at a value greater than 1,  b0 is the location parameter which must be greater than zero. Figure 1.2 EEG Decomposition and Extraction The Figure 1.2 shows the EEG decomposition and extraction using the wavelet transformation. As shown in the figure each of the extracted band components are sent through the “Windowing step”. In this step the interesting and boring portions of the bandcomponents basedonthetimestampsof the original EEG are extractedandsentthrougha windowing mechanism. In this mechanism, each bandcomponentsignal is partitioned into tiny windows.Afterthisstep eachwindow is decomposed using DWT. The decomposition process in wavelet transform can be performed iteratively into several levels. The values of the coefficients from the window interval is considered as the extracted features of EEG signals. 1.3 BRAIN WAVES CLASSIFICATION Brain patterns form wave shapes that are commonly sinusoidal. Usually, they are measured from peak to peak and normally range from 0.5 to 100 µV in amplitude, which is about 100 times lower than ECG signals. By means of Fourier transform powerspectrum fromtherawEEGsignal is derived. In power spectrum contribution of sine waves with different frequenciesarevisible.Althoughthespectrum is continuous, ranging from 0 Hz up to one half of sampling frequency, the brain state of the individual maymakecertain frequencies more dominant. Brain waves have been categorized into four basic groups: Figure 1.3 Different Categories of Brain Waves 2. SYSTEM DESIGN The schematic diagram illustrates the overall method ofthis project. As shown EEG signals areacquiredfromthesubjects during the experiment. In the project we examine 10 students who watch educational videos of two kinds (Confusing one and the one which is from an easy topic). The duration of the video is 2 minutes and we take voltage difference of the waveforms emitted by the brains of the students. Thus, the brain wave pattern is monitored using electrodes and sensors. After the activity of watching the video, students rate their level of confusion, either 1 or 0.’0’ means the topic is not confusing and the value ‘1’ means the topic was new or surprising to them. The entire video of 2 minutes is cut into 1 min (by removing the first 30s and last 30s). And the response fromstudentsis taken for each 5s session. Thus, getting 12 modes for one video for one user. We have two .csv files as our datasets in this project. The first one gives us the demographic information about the students and it is not much of use to us, since we perform our operations and apply our algorithms on the second file, which gives us the EEG signal information. We initially perform our description analysis on this dataset by viewing the number of rows and columns present in it and also checking for missing values if any. There are actually 15 columns out of which 4 of them(studentid,video id, user defined label, pre - defined label) are converted into integer for us to group and analyze the dataset easily.
  • 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 838 The 4 described columns will also not affect the understanding of the particular subject for the student. So, they are included in the process of feature extraction. Since, all the features are of numerical data type (floating point numbers), we apply Pearson’s coefficient as our tool. After applying, the Pearson’s coefficient we find the relationship between the features by plotting them in different plots (Cross tables, bar graphs and heatmaps). Later we apply Logistic Regression, RNN, Boost and Decision Tree on the same. Figure 2 Comparison of Bar Graphs We can infer from the above graph (Figure 3.a) that, video 9 was the easiest video since, only 2 students have told the video was surprising to them. In the same way, subject 1 has more learning capability than others since, he found only 4 videos were surprising. 2.1 DATA PREPROCESSING AND EXTRACTION The EEG signal is comprised of a complex and nonlinear combination of several distinct waveforms which are also called band components. Each of the band components is categorized by the frequency range that they exist in. The state of consciousness of the individuals may make one frequency range more pronounced than others. As shown in Figure above, the different band components are extracted from the raw EEG signal using Butterworth bandpass filters. Five primary bands of the EEG signal are extracted, namely, Delta (0.2–4Hz), Theta (4– 8Hz), Alpha (8–13Hz), Beta (13– 30Hz), and Gamma (30– 55Hz). This Section tells about how key features are extracted from raw EEG signal. This Sample EEG data is of two subjects. It’s given in a .EDF file which is basically Raw signal. The Output of such file is given below Figure 2.1 Raw EEG Signals The above .EDF file representation (Figure 2.1) has 7 signals in which 2 are EEG channels. 2.2 FEATURE SELECTION Feature selection is a process whereyouautomaticallyselect those features in your data that contribute most to the prediction variable or output in which you are interested. Types of feature Selection performed  Univariate Selection (operates only on positivevariables)  Recursive Feature Selection  Pearson Correlation  Feature Importance 2.2.1 UNIVARIATE SELECTION Univariate Selection is a set of Statistical tests to selectthose features that have the strongest relationship withtheoutput variable. Univariate Selection Works only on Positive Variable as we have a feature Raw which has negative values. The remaining 10 features are used for the classification. After Selecting the 10 features we use SelectKBest and chi2 to find the best features among the selected 10. Later the most Dominant Features are used for Classification using Decision Tree, Logistic Regression, Extreme Gradient Boosting. 2.2.2 RECURSIVE FEATURE SELECTION Recursive feature elimination is a feature selection method that fits a model and removes the weakest feature (or set of features) until the specified number of features is reached. This method works by recursively selecting attributes and building a model on those attributes that remain. After selecting the 10 features we use RFE to findthe bestfeatures
  • 4. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 839 among the selected 10. 2.2.3 PEARSON CORRELATION When a data set has too many variables. Most of the variables are correlated. So, running a model on whole data returns poor accuracy. Finding the most positively correlated variables is mostly the best choice. Because they affect the output the most. Correlation is the mutual relationship between two features. After Selecting the 11 features we use correlation (Pearson) to find the best features among the selected 11. Later the most Dominant Features are used for Classification using Decision Tree, Logistic Regression, Extreme Gradient Boosting. Figure 2.2.3 Pearson Correlation Using Pearson Correlation, the heat map looks like this (Figure 2.2.3) it shows Delta, Alpha, Gamma, Beta, Theta are strongly correlated and are most dominant. 2.2.4 FEATURE IMPORTANCE A common approach to eliminating features is to describe their relative importance to a model, then eliminate weak features or combinations of features andre-evaluatetoseeif the model fairs better during cross-validation this is feature importance. This method gives the feature importance of each feature of the dataset by using the feature importance property of the model.Featureimportancegivesthescorefor each feature of your data, the higher the score more important or relevant is the feature towards the output variable. After Selecting the 11 features we use ExtraTreesClassifier to find the best features among the selected 11. Later the most Dominant Features are used for Classification using Decision Tree, Logistic Regression, Extreme Gradient Boosting. 2.2.4 RAW SLEEP EEG DATASET The Sleep Physio net dataset is annotated using 8 labels <physionet_labels> Wake (W),Stage1,Stage2,Stage3, Stage 4 corresponding to the range from light sleep to deep sleep, REM sleep (R) where REM is the abbreviation for Rapid Eye Movement sleep, movement (M), and Stage (?) for any none scored segment. Figure 2.2.4 Sleep Stages We will work only with 5 stages: Wake (W), Stage 1, Stage 2, Stage 3/4, and REM sleep (R). Here Stage 1, Stage 2, Stage 3/4 they are nothing just the sleep concentration where Stage 1 has less sleep concentration than Stage 2. This is the Sleep data collected from a person (16 hrs. data). It has following values. Where FpZ (Intersection between Fz and Fp1 and Fp2 in the figure), Fz, Cz, Oz is just electrode labelling 2.2.5 DEPENDENT FEATURES(GRAPH) Here each abbreviation means. A-Attention, M-Meditation, R-Raw, D-Delta, A1-Alpha1, A2- Alpha2, B1-Beta1, B2-Beta2, G1-Gamma1, G2-Gamma2 The set of features dependent on each other is as follows: MAIN FEATURE DEPENDENT FEATURE R {A1, A2, B1, B2, G1, G2} {A1, A2, G1, D, M} A {G2, A, D} M
  • 5. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 840 From the above graph following inferences are made. A (attention) is most dependent among all, R (Raw) is the origin of all-important feature like alpha, gamma etc. Hence, we conclude Attention is the most dominant feature in our study. 3. RESULT After the code was completed various parameters were tuned for the best results. Adding all the features provided more accuracy than with using selected features. Classifiers were also changed to see which gives the best result. Here is the list of Classifier tried and the corresponding accuracies obtained with their corresponding run-time.Classificationis one of the methods used to identify the certain pattern from feature extraction such as amplitude of EEG data, power spectral density (PSD), and Band power (BP). The performances of identify the pattern are depend on both features and classification algorithm. There are many classification algorithms used in the EEG analysis. Here following classifiers are used. 3.1 COMPARISON  Better Dataset with high amount of training data.  Independent features at least one so that all features are not needed for classification.  Adding two new columns specifying the level of difficulty to make predictions more accurate. The Figure 3.1(a) shows 3 algorithms along with their run- time, clearly taking all features gives the best accuracy but it does take a lot of time. Also, you can see XGBoost takes more time than the other two because it keeps searching for local minima and then finds a common global minimum. Hence it takes more time than other two classifiers. Figure 3.1(a) Comparison of Classifiers Figure 3.1(b) RNN Accuracy The Figure 3.1(b) shows the accuracy of RNN. As seen after taking hidden layers greater than 10 the model goes to a saturation state and the accuracy remains same. 3. CONCLUSIONS Predicting the learning capabilities of an individual is not an easy task as it has lot of complex tasks involved. It can be seen that reasonable amount of accuracy achieved when all the features are selected. Amongst the various machine learning models investigated the classification using Extreme gradient boosting seems toperformbest(63.6%) on the combined feature set. However logistic regression using all features and recursive feature selection gives the closest accuracy to the best(61.3&, 61.2%).Also,byapplying various feature selection algorithms on the dataset wecome to know that we get the maximum accuracy of prediction when all the features are selected. This not only demonstrates that our dataset is small but also depicts that for making the prediction we can’t depend only on our four important features(theta,alpha,beta,gamma).Although this work is limited to EEG data there is a room for improvement using clinical data and genetic data. Here are some of the future work planned to improve the system’s classification and prediction performance.  A larger data set is needed to further validate this experiment. A larger data set is expected to provide a more robust classifier model.
  • 6. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 02 | Feb 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 841  Having a more diverse base of features usually provides insight into some connate characteristics of the signal which might not be openly evident. 
  Feature pruning and other classification methods need to be tried for increasing the accuracy. 
 Hence this project gives a little insight of how brain react to certain changes and how those changes can be used to predict certain tasks. REFERENCES [1] T. Blakely, K.J. Miller, R. P N Rao, Mark D.Holmes,andJ.G. Ojemann, “Localization and classification of phonemes using high spatial resolution electrocortico graphy(ECoG) grids,” in Engineering in Medicine and Biology Society, 2008. EMBS 2008. 30th Annual International Conference of the IEEE, Aug 2008, pp.4964–4967. [2] Spencer Kellis, Kai Miller, Kyle Thomson, Richard Brown, Paul House, and Bradley Greger, “Decoding spoken words using local field potentials recordedfrom thecortical surface,” Journal of Neural Engineering, vol. 7, no. 5, pp.1–10,2010. [3] Jess Bartels, Dinal Andreasen, Princewill Ehirim, HuiMao, Steven Seibert, E. Joe Wright, and Philip Kennedy,“Neurotrophic electrode: Method of assembly and implantation into human motor speech cortex,” Journal of Neuro science Methods, vol. 174, no. 2, pp.168–176,2008. [4] https://www.kdnuggets.com/2017/10/xgboost-top- machine-learning-method-kaggle-explaind.html [5] https://physionet.org/physiobank/database/sleep- edfx/sleep-cassette