EEG Emotion Prediction - Data Every Day
ฝัง
- เผยแพร่เมื่อ 25 ม.ค. 2025
- Hi guys, welcome back to Data Every Day!
On today's episode, we are looking at a dataset of EEG readings taken from subjects while they were watching various movie scenes and trying to predict their emotional states during the scenes. We will be using a TensorFlow neural network to make our predictions.
Here is a link to the Kaggle dataset:
www.kaggle.com...
And here is a link to my notebook from the video:
www.kaggle.com...
Thanks so much for watching! If you enjoyed today's episode, be sure to subscribe and hit the bell for more content!
Enjoy, and see you all tomorrow! :)
----------
Patreon: / gcdatkin
LinkedIn: / gcdatkin
Twitter: / gcdatkin
I learned more from watching you code this than being in a class for weeks. Thank you for so much for sharing your knowledge with the internet :)
DUDE! Mind blown! Love the pace and style! Thank you for sharing your process. I apparently have a lot of catching up on quality content to do... and digging for pennies for your patreon.
Had lectures from physics through medicine and back and never encountered someone so clear and concise in their explanation. Hat's off.... still collecting my brain off the floor.
oh and the lack of "shine" and multiple intros is deeply refreshing!
Well done Gabriel Atkin, I am learning through your code.
Keep it up.
This channel is going to blow up one day! Keep up the great work :)
Thanks so much! I hope so :)
Great video! Thanks for using our dataset :)
No problem! Really nice dataset :)
Hi Jordan, thanks for this awesome dataset. I was wondering what the features from this set specifically are? In the video, Gabiel works with the assumption that it's all time series, but I was wondering if this is the case. Thanks!
@@sarahn6292 If you check Jordan Bird’s youtube channel, there is a video about how the dataset was produced, which largely explains what the dimensions in the dataset are. Furthermore, in the video description for that video, there is a link to the code used to produce the dataset from the EEG data as output by a particular model of EEG device, and that, I imagine, should clarify any remaining potential ambiguity.
(I haven’t read through said code.)
My understanding of the video is that it is time series data, sorta, but that essentially uh,
well, you know how in this video he said that the time series data often has 3 indices, one for the “which sample”, one for “what timestep”, and one for “what feature at that timestep”?
Well, my impression is that essentially those last two are sorta flattened into one index here, and that is what the different columns are.
I’ve not done any real ML stuff, but my guess is that by grouping together the columns with the same number, in order to make the three-dimensional array, uh, that would better interpret the dataset as time series than how this video shows? (Though given how high accuracy this video managed, seems to suggest that I may have misunderstood something.)
But, like, check the actual source, not just my summary of it, as I am likely to have gotten details wrong, or gotten mixed up in other ways.
Thank you so much!
You're explanation is amazing! I am really grateful for you .
You have no idea how much your video helped me in my project.
Keep up the good work.
I just wanted to get a brief idea about this topic for my final year project ... And now i am just awestruck with thr content of this channel and the videos! Keep it up! 🙏
Thanks so much Soumita!
I'm working on seed-IV dataset by using EEG signal and will follow your algorithms in this video.
Thank you brother.
Awesome video thanks for sharing this great tutorial. Really you doing great effort.
No problem!
Hi bro
I have a doubt,
How to predict the emotions like sad,happy,angry etc by using this method?
Please reply bro🙏
Could you please check the Notebook link it doesn't work for me. Thanks
Oh, sorry! I forgot to make it public. It should work now.
@@gcdatkin thanks very much
A KerasTensor cannot be used as input to a TensorFlow function. A KerasTensor is a symbolic placeholder for a shape and dtype, used when constructing Keras Functional models or Keras Functions. This error is coming at
expand_dims = tf.expand_dims(inputs, axis=2)
gru = tf.keras.layers.GRU(256, return_sequences=True)(expand_dims)
hi! How did you preprocessed your data? I mean did you use statistics?
Are there any data sets for MATLAB?
fft column could stand for fast fourier transform
Thanks a lot
Hey, really liked this vid. Is there any way to get in touch with who created the data set or something to get more details on what the features mean? I just feel like that's an important part of making sense of this data.
It looks like the dataset's author also commented on this video (Jordan). Maybe you could reply to his comment with a question!
@@gcdatkin Thanks!
Where can i find dataset ur using
I always include a link in the video description. Here is the link for this video:
www.kaggle.com/birdy654/eeg-brainwave-dataset-feeling-emotions
Sir that data is in featured format ...? I have EEG signals in text document for classifying emotions ..so how can I convert my data as yours ...I mean as fft ....? Pleased do help me
Ur videos are helping but am finding difficult in converting my data as urs .. please do help me for that
awesomeee, thank youu
Sir,pls im requesting u
Can u pls send some other dataset csv file which works for this code??
awesome sir
love you bro
cool !