How to recognize emotions by electroencephalogram?

Question

Machine learning

Hello.
There are several files with electroencephalogram data. Each file is a record of one person (a matrix, where the columns are the numbers of sensors and the column with the numbers of emotions, and the rows are the numbers of measurements). It turns out that the file consists of the readings of the electroencephalogram sensors and the emotion number for each measurement. Basically, there is only one emotion per file, but there are files where the emotion changes (for example, from a neutral or sad state to a joyful one).
A tool is required that can determine human emotions based on electroencephalogram data (not necessarily neural networks).
Thank you)

Answer 1

Neural networks - when learning, they look for patterns. I think you can train a neural network to recognize emotions - but you need to take a larger sample (not just one person). An encephalogram is a combination of the activity of the brain regions - in principle, the types of activity with the same emotions should match in the same people (people with approximately the same life experience). But for correct operation, a large amount of incoming data is required - you need to take at least 10,000 people and record encephalograms for a month with decoding of emotions by points in time.
In general, this is a topic for a dissertation, and not for a question on a toaster.
Most likely, you will not succeed with your data - since they are abstract from a statistical point of view (there are very few of them).